01-09 Stack Configuration
01-09 Stack Configuration
9 Stack Configuration
This chapter describes how to set up a stack of multiple switches to improve forwarding
performance and reliability.
To check whether a switch supports service port stacking using dedicated stack cables, perform the
following operations: Select Service port stacking, and then check whether you can select 0.5m and
1.5m SFP+ dedicated stack cable or 0.5m and 1.5m SFP+ dedicated stack cable (supported from
V200R011C10) as stack cables. If so, switches of this model support service port stacking using
dedicated stack cables.
2. Click Query after finishing the selection. The stack precautions, connection rules, and
software configuration are displayed. Set up a stack according to these information.
Definition
Intelligent Stack is a technology that enables multiple stacking-capable switches to function
as a single logical switching device. This technology is also called iStack. As shown in
Figure 9-3, after SwitchA and SwitchB are connected by stack cables and form a stack, their
upstream and downstream devices consider them as one switch.
Network Network
Stack
Equivalent to
SwitchA SwitchB Switch
Stack link
Purpose
Stack technology provides high network reliability and scalability, while simplifying network
management.
l High reliability: Member switches in a stack work in redundancy mode. Link
redundancy can also be implemented between member switches through inter-device
link aggregation. Member switches can join or leave the stack without affecting other
member switches. New switches automatically synchronize the configuration file and
system software version with the master switch.
l High scalability: You can increase ports, bandwidth, and processing capacity of a stack
by simply adding member switches to the stack.
l Simplified configuration and management: You can log in to a stack from any member
switch to manage and configure all the member switches in the stack. In addition,
complicated Layer 2 ring protection protocols or Layer 3 protection switching protocols
are not required after switches set up a stack, making the network configuration much
simpler.
l Role
Switches that have joined a stack are member switches. Each member switch in a stack
plays one of the following roles:
– Master switch
The master switch manages the entire stack. A stack has only one master switch.
– Standby switch
The standby switch is the backup of the master switch. When the master switch
fails, the standby switch takes over all services from the master switch. A stack has
only one standby switch.
– Slave switch
A slave switch forwards service traffic. The more slave switches in a stack, the
higher forwarding performance the stack can provide. Apart from the master and
standby switches, all the other switches in a stack are slave switches.
l Stack ID
The stack ID of a member switch is the slot ID of the member switch within the logical
switch. Each member switch in a stack must have a unique stack ID.
l Stack priority
The stack priority of a member switch determines the role of the member switch in role
election. A larger value indicates a higher priority and higher probability that the
member switch will be elected as the master switch.
Master
Stack
Stack
Standby Slave
Ring topology l High reliability: The first and last A ring topology is
If a stack link member switches recommended when
fails, the need to be connected member switches
topology changes by a physical link, are located near one
from ring to so this topology is another, because this
chain, and the not appropriate for topology has higher
stack can still long-distance reliability and link
function stacking. utilization.
normally.
l High link
bandwidth
efficiency: Data
can be forwarded
along the shortest
path.
Master Election
Determine the stack connection mode and topology, connect the member switches with
physical links, and then power on all member switches. These member switches elect the
master switch, which manages the stack. The master switch is elected based on the following
rules (the election ends when a winning switch is found):
The master election timeout interval is 20 seconds. The startup process may take
different lengths of time on different member switches. When stack member switches are
powered on or restart, some member switches may not participate in the first master
election. When a switch that starts later joins the stack, the master switch is elected
again. If the previous master switch fails the election, it restarts and then joins the stack
as a non-master switch. If the switch that starts later fails the election, it can join the
stack only as a non-master switch. For details, see 9.3.5 Joining and Leaving a Stack.
If you want a specific switch to act as the master switch, power on that switch first, and
power on the other switches after this switch starts.
NOTE
To ensure that master election is completed at a time, you are advised to use switches of the same model
to set up a stack. If you want to set up a stack of different switch models, you are advised to connect
switches of the same model together.
For example, three switches A, B, and C set up a stack in a chain topology.
– If A and B start first and C starts later, C joins the stack only as a non-master
switch.
– If A and C start first and become the master switches, A and C compete to be the
master switch based on their startup time when B starts and joins the stack. The
switch that fails the election restarts and joins the stack as a non-master switch.
For example, four switches A, B, C, and D set up a stack in ring topology:
– If A and B start first and C and D start later, C and D join the stack only as non-
master switches.
– If A and C start first and become the master switches, A and C compete to be the
master switch based on their startup time when B and D start and join the stack. The
switch that fails the election restarts and joins the stack as a non-master switch.
2. If multiple switches complete startup at the same time, the switch with the highest stack
priority becomes the master switch.
3. If multiple switches complete startup at the same time and have the same stack priority,
the switch with the smallest MAC address becomes the master switch.
Running
After role election and topology information collection are complete, all the member switches
synchronize their system software and configuration files with the master switch.
l Automatic software loading: Member switches can run different software versions. If
different software versions running on the member switches are compatible with one
another, the member switches can set up a stack. If the software version running on the
master switch is different from those on the standby switch and slave switches, these
switches download the system software from the master switch. Then, they restart with
the new system software and rejoin the stack.
l Configuration file synchronization: The standby switch and slave switches can download
the configuration file of the master switch and apply this configuration file. This
mechanism enables member switches to work like a single device and ensures that other
switches continue working normally if the master switch fails.
A stack Initial stack Configure stack IDs for Applicable and Applicable and
is newly IDs of all member switches one by recommended recommended
set up. member one before setting up a
switches are 0. stack.
1. Place member
devices.
2. Set the stack IDs of
member switches one
by one to the expected
values.
3. Connect member
switches using stack
cables to set up a
stack.
Initial stack Configure stack IDs for Applicable and Applicable and
IDs of some member switches one by recommended recommended
member one before setting up a
switches are stack.
not 0. 1. Place member
devices.
2. Set the stack IDs of
member switches one
by one to the expected
values.
3. Connect member
switches using stack
cables to set up a
stack.
The Stack IDs of Connect new member Applicable and Applicable and
stack new member switches using cables and recommended recommended
capacity switches are 0. then power on these
needs to switches.
be 1. Place member
expande devices.
d.
2. Connect the new
member switches to
the stack using cables.
3. Power on the new
member switches one
by one.
Stack IDs of Configure stack IDs for Applicable and Applicable and
new member new member switches recommended recommended
switches are one by one before
not 0. capacity expansion.
1. Place member
devices.
2. Set the stack IDs of
the new member
switches one by one to
the expected values.
3. Connect the new
member switches to
the stack using stack
cables.
A Log in to the member switch to clear the startup configuration file and restore
member its stack ID to 0.
switch
needs to
be
replaced
.
A stack Log in to the member switch to be removed to clear the startup configuration
needs to file and restore its stack ID to 0.
be split.
A stack of multiple switches acts as a virtual device. The interface numbering rules, system
login methods, and file system access methods used in the stack are different from those used
on standalone switches.
After a switch joins a stack, its interfaces are numbered in the stack ID/subcard ID/port
sequence number format. In a stack, each member switch has a unique stack ID.
The subcard ID and port sequence number remain unchanged after a switch joins a stack.
After a stack member switch leaves the stack, it still uses the stack ID as its slot ID.
NOTE
The interface number of a management interface is fixed as MEth0/0/1 regardless of whether the switch joins
a stack or what the stack ID is. In a stack, only the management interface on one member switch takes effect
and becomes the master management interface.
Stack Login
You can log in to a stack using the following methods:
l Local login: Log in through the console interface of any member switch.
l Remote login: Log in through the management interface or another Layer 3 interface of
any member switch. You can remotely log in to the stack using Telnet, STelnet, web, or
SNMP if your operation terminal has a reachable route to the stack.
NOTE
l After switches with management interfaces form a stack, only the management interface on
one member switch takes effect and becomes the master management interface. After the stack
starts, the management interface on the master switch is used as the master management
interface. If the management interface on the master switch is faulty or unavailable, the
management interface on another member switch is used as the master management interface.
If you connect your PC directly to a non-master management interface, you cannot log in to
the stack.
l After a stack is set up, all the member switches use the configuration file of the master switch.
l A stack externally acts as a single switch and uses one management IP address. To remotely
log in to a stack, you need to log in to the IP address of the master switch.
l All stack member switches are able to forward traffic.
After you log in to the stack, the master switch issues the configurations to the other member
switches. This allows resources of member switches to be managed consistently.
l A switch can be added to a stack while it is powered on or off. This section describes how a member
switch joins a stack after being powered off. For details on how a member switch joins a stack after
being powered on, see Stack Merging.
l It is not recommended to add a member switch to a stack while the power is on.
SwitchA
(Master)
SwitchD
Stack +
SwitchB SwitchC
(Standby) (Slave)
SwitchD joins
SwitchA SwitchD
(Master) (Slave)
Stack
SwitchB SwitchC
(Standby) (Slave)
switch is elected between the two original master switches. The master switch in the stack that
enters the running state first becomes the new master switch. If the two stacks enter the
running state at the same time, the master switch is elected based the same rules used when a
stack is set up.
SwitchA
(Master)
SwitchD SwitchE
(Master) (Standby)
Stack1 + Stack2
SwitchB SwitchC
(Standby) (Slave)
SwitchA SwitchD
(Master) (Slave)
Stack
SwitchB SwitchE
(Standby) (Slave)
SwitchC
(Slave)
Stack Split
If you remove some member switches from a running stack without powering off the switches
or if multiple stack cables fail, the stack splits into multiple stacks.
Depending on the locations of the master and standby switches after a split, device roles are
elected in either of the following ways:
l If the original master and standby switches are in the same stack after the split, the
master switch calculates the stack topology by deleting topology information related to
the removed member switches, and then synchronizes updated topology information to
the other member switches. When removed member switches detect that the timeout
timer for stack protocol packets has expired, the switches restart and begin a new master
election.
As shown in Figure 9-7, the original master switch (SwitchA) and standby switch
(SwitchB) are in the same stack after the split. SwitchA deletes topology information
related to SwitchD and SwitchE and synchronizes topology information to SwitchB and
SwitchC. After SwitchD and SwitchE restart, they set up a new stack.
Figure 9-7 Original master and standby switches in the same stack after a split
SwitchA SwitchD
(Master) (Slave)
Stack
SwitchB SwitchE
(Standby) (Slave)
SwitchC
(Slave)
Stack splits
SwitchA and SwitchB
are in the same stack
after the split
Stack1 Stack2
+
SwitchA SwitchB SwitchC SwitchD SwitchE
(Master) (Standby) (Slave) (Master) (Standby)
l If the original master and standby switches are in different stacks after the split, the
original master switch selects a new standby switch in its stack, calculates stack topology
information, and synchronizes updated topology information to the other member
switches. The original standby switch becomes the new master switch in its stack, and
then calculates stack topology information, synchronizes stack topology information to
other the member switches, and selects a new standby switch.
As shown in Figure 9-8, the original master switch (SwitchA) and standby switch
(SwitchB) are in different stacks after the split. SwitchA specifies SwitchD as the new
standby switch, calculates stack topology information, and synchronizes topology
information to SwitchD and SwitchE. In the other stack, SwitchB becomes the master
switch. SwitchB then calculates topology information, synchronizes topology
information to SwitchC, and specifies SwitchC as the new standby switch.
Figure 9-8 Original master and standby switches in different stacks after a split
SwitchA SwitchD
(Master) (Slave)
Stack
SwitchB SwitchE
(Standby) (Slave)
SwitchC
(Slave)
Stack splits
SwitchA and SwitchB
are in different stacks
after the split
Stack1 Stack2
+
SwitchB SwitchC SwitchA SwitchD SwitchE
(Master) (Standby) (Master) (Standby) (Slave)
Multi-Active Detection
All member switches in a stack use the same IP address and MAC address (that is, the MAC
address of the stack). Therefore, after a stack splits, more than one stack may use the same IP
address and MAC address. To prevent this situation, a mechanism is required to check for IP
address and MAC address collision after a split.
Multi-active detection (MAD) is a stack split detection protocol. When a stack splits due to a
link failure, MAD provides split detection, multi-active handling, and fault recovery
mechanisms to minimize the impact of a stack split on services.
MAD can be implemented in direct or relay mode. The direct and relay modes cannot be
configured together in the same stack.
l Direct mode
In direct mode, stack member switches use direct links over common network cables as
dedicated MAD links. When the stack is running normally, member switches do not send
MAD packets. After the stack splits, the member switches send a MAD packet every 1s
over the MAD link to check whether more than one master switch exists.
In direct mode, stack member switches can be directly connected to an intermediate
device or fully meshed to each other:
– Directly connected to an intermediate device (Figure 9-9): Each member switch has
at least one MAD link connected to the intermediate device.
– Fully meshed to each other (Figure 9-10): In the full-mesh topology, at least one
MAD link exists between any two member switches.
The use of an intermediate device can shorten the MAD links between member switches.
This topology applies to stacks where member switches are far from each other. The full-
mesh topology prevents MAD failures caused by failures of an intermediate device, but
full-mesh connections occupy many interfaces on the member switches. Therefore, this
topology applies to stacks with a few member switches.
NOTE
l After configuring MAD in direct mode on an interface, do not configure other services on the
interface.
l A maximum of eight direct MAD links can be configured between member switches to ensure
reliability.
l MAD packets are bridge protocol data units (BPDUs), so the intermediate device must be able
to forward BPDUs. For details on how to configure this function, see Configuring Interface-
based Layer 2 Protocol Transparent Transmission in "Layer 2 Protocol Transparent
Transmission Configuration" in the S1720, S2700, S5700, and S6720 V200R011C10
Configuration Guide - Ethernet Switching.
SwitchD
Stack
MAD Link
Stack Link
Stack
MAD Link
Stack Link
l Relay mode
In relay mode, MAD relay detection is configured on an Eth-Trunk interface in the stack,
and the MAD detection function is enabled on an agent. Every member switch must have
a link to the agent and these links must be added to the same Eth-Trunk. In contrast to
the direct mode, the relay mode does not require additional interfaces because the Eth-
Trunk interface can run other services while performing MAD relay detection.
In relay mode, when the stack is running normally, member switches send MAD packets
at an interval of 30s over the MAD links and do not process received MAD packets.
After the stack splits, member switches send MAD packets at an interval of 1s over the
MAD links to check whether more than one master switch exists.
You can use an independent relay agent (Figure 9-11) or use two stacks as each other's
relay agents (Figure 9-12).
NOTE
l The relay agent is a switch that supports the MAD relay function. Currently, all the S series
switches support this function.
l To implement MAD relay detection by using two stacks as each other's relay agent, configure
different domain IDs for the two stacks. Member switches of a stack form a stack domain. A
network may have multiple stack domains, with different domain IDs.
SwitchD
Eth-Trunk
Stack
MAD Link
Stack Link
Stack
Domain ID 1 SwitchA SwitchB SwitchC
Eth-Trunk
MAD Link
Stack Link
Each Eth-Trunk can have a maximum of eight member interfaces. Therefore, when a
stack contains nine member switches, one Eth-Trunk cannot provide MAD links for all
the member switches. In this case, configure multiple Eth-Trunks to ensure that a MAD
link is available between every two member switches. As shown in Figure 9-13, Eth-
Trunk1 provides MAD links for Switch1 through Switch8, Eth-Trunk2 provides MAD
links for Switch2 through Switch9, and Eth-Trunk3 provides MAD links for Switch1 and
Switch9.
Eth-Trunk1
Eth-Trunk2
Eth-Trunk3
Eth-Trunk3
Stack
MAD Link
Stack Link
Multi-Active Handling
After a stack splits, the MAD mechanism sets the new stacks to Detect or Recovery state. The
stack in Detect state still works, whereas the stack in Recovery state is disabled.
MAD handles a multi-active situation in the following way: When multiple stacks in Detect
state are detected by the MAD split detection mechanism, the stacks compete to retain the
Detect state. The stacks that fail in the competition enter the Recovery state, and all the
physical ports except the excluded ports on the member switches in these stacks are shut
down. The stacks in Recovery state no longer forward service packets.
Fault Recovery
After the faulty link recovers, the stacks merge into one in either of the following ways:
l The stack in Recovery state restarts and merges with the stack in Detect state, and the
service ports that have been shut down are restored to Up state. Then the entire stack
recovers.
l If the stack in Detect state is also faulty before the faulty link recovers, you can remove
this stack from the network and start the stack in Recovery state using a command to
switch service traffic to this stack. Then rectify the stack system fault and link fault.
After the stack recovers, connect it to the network so that it can merge with the other
stack.
SwitchA SwitchA
(Master) (Slave)
Switchover
Stack1 Stack2
Intelligent Upgrade
When a stack is set up or new member switches join a stack, the standby and slave switches or
new member switches check whether their software versions are the same as that of the
master switch. If not, they download the system software from the master switch, restart with
the new system software, and rejoin the stack.
Traditional Upgrade
You can specify the startup system software on the master switch and restart the entire stack
to upgrade the software of member switches. This upgrade method causes service interruption
for a relatively long time and can be used in scenarios insensitive to the service interruption
time.
Smooth Upgrade
A smooth upgrade can be performed in a stack that has uplinks and downlinks working in
redundancy mode. As shown in Figure 9-15, the stack is divided into an active area (with the
master switch) and a backup area. Member switches in the two areas are upgraded in turn.
When an area is being upgraded, traffic is transmitted through the other area, minimizing the
impact of the upgrade on services. Use this upgrade method for the scenarios that are sensitive
to the service interruption time.
In a network where uplinks and downlinks work in redundancy mode, both upstream traffic
and downstream traffic are forwarded through the active and backup areas. When member
switches in the backup area are upgrading, traffic is forwarded through the active area. When
member switches in the active area are being upgraded, traffic is forwarded through the
backup area.
NOTE
To implement smooth upgrade for a stack, the stack must have uplinks and downlinks working in
redundancy mode.To ensure a successful upgrade, you are advised to add at least one link on each
member switch to an inter-device Eth-Trunk.
Network
Building A Building B
Stack link
common link
Eth-Trunk
Active area
2 1 Master
4 3
6 5
Backup area
If errors or exceptions occur during a smooth upgrade, or the upgrade times out (the timeout interval is
70 minutes), member switches automatically roll back to the original system version and set up a stack
again.
fails, traffic on this link is distributed to the other link through the stack cables connecting the
member switches. This link backup mechanism improves network reliability.
Network Network
Stack Stack
Data flow 1
Data flow 2
Stack Link
Eth-Trunk
As shown in Figure 9-18, traffic sent to the core device on the network is equally distributed
to member links of an Eth-Trunk set up between stack member switches. When a member
switch fails, traffic toward this switch is distributed to the other switch. This device backup
mechanism improves network reliability.
Network Network
Stack Stack
As shown in Figure 9-19, SwitchA and SwitchB form a stack, and their uplink and downlink
interfaces are bundled to Eth-Trunk interfaces. If local preferential forwarding is not
configured, traffic reaching SwitchA is load balanced between the Eth-Trunk member links.
Some traffic is forwarded from SwitchA to SwitchB through the stack cables and is sent out
from a physical interface on SwitchB. If local preferential forwarding is configured, traffic
reaching SwitchA is forwarded through a local physical interface and does not pass through
the stack cables. By default, the local preferential forwarding function is enabled on the
device.
Network Network
SwitchC SwitchC
Increasing Bandwidth
As shown in Figure 9-21, when higher uplink bandwidth is required, you can add new
member switches to the stack and bundle physical links of the member switches into a link
aggregation group to increase the uplink bandwidth.
Stack
MSTP + VRRP
Stack
Long-Distance Stacking
As shown in Figure 9-23, users on each floor connect to the external network through
respective corridor switches, which are far from one another. When corridor switches in a
building are connected using stack cables and form a stack, each building has only one virtual
access switch, so the network becomes much simpler. Furthermore, each building has
multiple links to the core network, improving network robustness and reliability. The
administrator only needs to configure stacks and does not need to configure all the corridor
switches one by one, reducing management and maintenance costs.
Network
Stack Stack
NOTE
Switches can set up a stack through stack card connection and service port connection based
on stack port types.
l Member switches are connected using dedicated stack cards ES5D21VST000 and stack
cables.
l Stack cards are integrated into the rear panels of member switches, and member switches
are connected using integrated stack ports and dedicated stack cables.
Stack Link
Maintain the stack You can perform the 9.13 Maintaining Stacks
following tasks during stack
maintenance:
l 9.13.1 Monitoring the
Stack Status
l 9.13.3 Performing a
Master/Standby
Switchover
l 9.13.4 Upgrading Stack
Software
The preceding tasks are
optional and can be
configured based on your
needs.
Split the stack If the stack is not required, 9.14 Splitting a Stack
split the stack to restore the
member switches to
standalone switches.
The following switch models can set up a stack system in this scenario: S6700EI, S5700HI,
S5710HI, S5710EI, S5700EI, S5700SI, S5720EI, S5720HI, S6720S-EI, and S6720EI.
In this scenario, each switch in a stack connects to a core device through Eth-Trunk. The stack
system simplifies management of aggregation devices and improves uplink reliability of
aggregation devices.
Aggregation Stack
Access
AP
User
The following switch models can set up a stack system in this scenario: S2720EI, S2750EI,
S5700LI, S5700EI, S5710-C-LI, S5710-X-LI, S5720LI, S5720S-LI, S5700SI, S5720SI,
S5720S-SI, S5700S-LI, S5730SI, S5730S-EI, S6720LI, S6720S-LI, S6720SI, and S6720S-SI.
In this scenario, each switch in a stack connects to an aggregation device through Eth-Trunk.
The stack system simplifies management of and improves uplink reliability of access devices.
Core CSS
AC
Aggregation
Stack
Access
AP
User
Aggregation
Access
Stack1 Stack2
Stack3
User AP
Recommended Scenarios
NOTE
The following recommendations are provided based on the positioning of fixed switch models. If customers
have special requirements, deploying high-end devices at a lower network layer is recommended; you are not
advised to deploy low-end devices at a higher network layer. For example, it is recommended to deploy
aggregation switches at the access layer rather than to deploy access switches at the aggregation layer.
To ensure stack reliability and bandwidth, you are advised to do as follows:
l Ensure that each member device connects to the core device through an uplink port. This connection
prevents upstream traffic forwarding from being affected when any member device fails.
l When using multiple devices to set up a stack, ensure the same stack bandwidth between any two
devices. Otherwise, the bandwidth of the stack system is the minimum stack bandwidth.
Feature Limitations
Stack Setup Guidelines
l A stack contains a maximum of five, eight, or nine stack members depending on switch
models. To ensure high forwarding performance and reliability, ensure that the number
of stack members of each switch series does not exceed the recommended value. For
example, the recommended number of S5700LI switches in a stack is 2 to 5. For the
recommended value of each switch series, see "Service Port Stacking Support" or "Stack
Card Stacking Support".
l When setting up a stack containing a mix of different switch models, follow the
following rules:
– A stack cannot be set up among different switch series. For example, a stack cannot
be set up among the S5700 and S6700 switches.
– A stack cannot be set up among different switch models of the same switch series.
For example, a stack cannot be set up among the S5720EI and S5720HI switches.
– A stack can be set up among only certain different switch models of the same
switch series. For example, a stack can be set up among the S5720SI and S5720S-
SI switches but not among the S5720LI and S5720S-LI switches. For details about
whether a stack can be set up among different switch models, see stack support of
different switch models in Stacking Support and Version Requirements. As
described in Remarks of S6720LI and S6720S-LI Service Port Stacking
Support, a stack can be set up among all S6720LI models, among all S6720S-LI
models, but not among the S6720LI and S6720S-LI models.
Version restrictions:
l When multiple switches set up a stack, member switches will synchronize the running
version of the master switch. If a member switch does not support this running version, it
will restart repeatedly.
l In V200R009C00, if MPLS-incapable S5720EIs exist in a stack, this stack cannot have
MPLS enabled. If member devices in a stack are running MPLS services, adding MPLS-
incapable S5720EIs to the stack is not allowed.
MAD specifications:
l You can configure a maximum of eight direct detection links for each member switch in
a stack.
l You can configure the relay mode on a maximum of four Eth-Trunks in a stack.
l In V200R008C00 and earlier versions, you can configure a maximum of 64 Eth-Trunks
on a relay agent to provide the relay function for multiple stacks. This restriction does
not apply to versions later than V200R008C00.
After multiple switches form a stack, the following features cannot be configured in the
stack:
When you establish a stack on the switches that support both stack card connection and
service port connection, such as S5720-C-EI, note the following:
l All member switches must use the same stack connection mode.
l When a member switch has stack cards installed and the service port stack configuration,
the switch uses the service port connection mode to establish a stack. It does not use the
stack card connection mode even though a stack fails to be established in service port
connection mode and stack cards are connected correctly.
l A switch uses the stack card connection mode to establish a stack only when it has no
service port stack configuration.
l If a switch is currently using the stack card connection mode, perform the service port
stack configuration on the switch before changing the stack connection mode to service
port connection. After the service port stack configuration is complete, the switch uses
the service port connection mode when restarting.
l If a switch using the stack card connection mode has service port configuration, a
smooth upgrade cannot be performed on the switch.
l If a switch is currently using the service port connection mode, correctly connect stack
cards and stack cables and clear the existing service port stack configuration before
changing the stack connection mode to stack card connection. You can use the reset
stack-port configuration command to clear the existing service port stack
configuration.
l When changing service port connection to stack card connection, you are advised to
remove the cables connected to service ports to prevent loops.
Deployment Recommendations
l Connect a stack to other network devices using an Eth-Trunk and add one port of each
member switch to the Eth-Trunk.
l When a stack connects to access devices, configure ports directly connected to terminals
as STP edge ports to prevent STP re-calculation when the ports alternate between Up and
Down states. This configuration ensures normal traffic forwarding.
l If storm control needs to be configured on many ports, replace storm control with traffic
suppression to save CPU resources.
l If port security needs to be configured on many ports, replace port security with MAC
address learning limiting to save CPU resources.
l Loops may occur on a network to which a stack connects. Run the mac-address
flapping action error-down command to set an interface to the error-down state when
MAC address flapping is detected on the interface. This improves system processing
performance and allows the peer device to detect that the interface becomes Down.
Additionally, if the peer device has redundant links, traffic can be rapidly switched to a
normal link.
Licensing Requirements
Stacking is not under license control.
Version Requirements
NOTE
Switches can set up a stack through stack card connection and service port connection based on stack port
types. Service port connections are classified into ordinary and dedicated cable connections based on cable
types.
l Ordinary cable connection: Switches use optical cables, network cables, and high-speed cables to set
up a stack.
l Dedicated cable connection: Switches use dedicated stack cables to set up a stack. This connection
mode is also called zero-configuration stacking.
For stacking support of switch models, see "Service Port Stacking Support" or "Stack Card Stacking
Support".
NOTE
To know details about software mappings, see Hardware Query Tool.
l For the cables that can be used as stack cables, see "Cables" in the Hardware Description. Using other
cables may cause exceptions in the stack system, for example, frequent error-down events on stack ports.
l Combo ports cannot function as stack member ports.
l For stacking-supported service ports, some of them can be used as stack ports, whereas some can be used
as ordinary service ports.
l After a UCL group is created using the ucl-group command on S2720EI switches, a stack cannot be set
up.
l Since V200R011C10, S2720EI and S2750EI series switches can set up stacks in service port stacking
mode using dedicated stack cables.
l The number of S2750EIs used to set up a stack is recommended to be 2 to 5.
l For S2700EIs and S2710SIs, if the stack system works only in Layer 2 and Layer 3 forwarding scenarios,
the number of stack member switches is recommended to be 2 to 5. If the stack system works in user
access or gateway scenarios that require protocol interaction, the number of stack member switches is
recommended to be 2 or 3.
l The S2720EI has only one model, S2720-28TP-EI-AC, in V200R006C10, V200R009C00, and
V200R010C00. The S2720-28TP-EI-AC cannot be upgraded to V200R011C00 and later versions. The
number of S2720-28TP-EI-ACs used to set up a stack is recommended to be 2 to 5. In V200R011C10 and
later versions, the number of S2720EIs used to set up a stack is recommended to be 2 to 9.
S270 First two GE 1.5 m SFP 2.5Gb l Among S2700EI switches, only the
0EI ports on the stack cable it/s S2700-52P-EI-AC and S2700-52P-
front panel PWR-EI support the stacking
function. A stack can be set up
between the S2700-52P-EI-AC and
S2700-52P-PWR-EI.
l A switch supports a maximum of two
logical stack ports, and each logical
stack port contains one stack member
port. A switch supports a maximum
of two stack member ports.
S271 First two GE 1.5 m SFP 2.5Gb l Any models of the S2710SI series
0SI ports on the stack cable it/s can set up a stack.
front panel l A switch supports a maximum of two
logical stack ports, and each logical
stack port contains one stack member
port. A switch supports a maximum
of two stack member ports.
S275 Two SFP l 1 m SFP+ 2.5 l Any models of the S2750EI series
0EI optical ports passive Gbit/s can set up a stack.
high- l On an S2750EI switch, only the first
speed and second service ports from the
copper right can be configured as stack
cables member ports of a logical stack port.
l 10 m SFP l A switch supports a maximum of two
+ active logical stack ports, and each logical
high- stack port contains one or two stack
speed member ports. Each switch can use a
copper maximum of two service ports as
cables stack member ports.
l 3 m and
10 m
AOC
cables
l 6GE stack
optical
module
(SFP-6GE
-LR) and
optical
fiber
l 0.5 m and
1.5 m SFP
+
dedicated
stack
cable
(supporte
d in
V200R01
1C10 and
later
versions)
l For the cables that can be used as stack cables, see "Cables" in the Hardware Description. Using other
cables may cause exceptions in the stack system, for example, frequent error-down events on stack ports.
l Combo ports cannot function as stack member ports.
l For stacking-supported service ports, some of them can be used as stack ports, whereas some can be used
as ordinary service ports.
l If the stack system works only in Layer 2 and Layer 3 forwarding scenarios, the number of stack member
switches is recommended to be 2 to 5. If the stack system works in user access or gateway scenarios that
require protocol interaction, the number of stack member switches is recommended to be 2 or 3.
S370 First two GE 1.5 m SFP 2.5Gb l All models support the stacking
0EI ports on the stack cable it/s function except that a stack cannot be
front panel set up between the S3752EI and
S3728EI.
l A switch supports a maximum of two
logical stack ports, and each logical
stack port contains one stack member
port. A switch supports a maximum
of two stack member ports.
S370 First two GE 1.5 m SFP 2.5Gb l All models support the stacking
0SI ports on the stack cable it/s function except that a stack cannot be
front panel set up between the S3752SI and
S3728SI.
l A switch supports a maximum of two
logical stack ports, and each logical
stack port contains one stack member
port. A switch supports a maximum
of two stack member ports.
l For the cables that can be used as stack cables, see "Cables" in the Hardware Description. Using other
cables may cause exceptions in the stack system, for example, frequent error-down events on stack ports.
l The number of stacked switches is recommended to be 2 to 5.
l For the cables that can be used as stack cables, see "Cables" in the Hardware Description. Using
other cables may cause exceptions in the stack system, for example, frequent error-down events on
stack ports.
l Combo ports cannot function as stack member ports.
l For stacking-supported service ports, some of them can be used as stack ports, whereas some can be
used as ordinary service ports.
l S5700LI series switches are classified into S5700-P-LI, S5700-TP-LI, and S5700-X-LI based on
their uplink interface types. A stack cannot be set up between S5700-P-LI, S5700-TP-LI, and
S5700-X-LI.
l S5700S-LI series switches are classified into S5700S-P-LI and S5700S-X-LI based on their uplink
interface types. A stack cannot be set up between S5700S-P-LI and S5700S-X-LI.
l Among S5700-P-LI series switches, only the S5700-28P-LI-AC, S5700-28P-LI-DC, S5700-28P-
PWR-LI-AC, S5700-52P-LI-AC, S5700-52P-LI-DC, and S5700-52P-PWR-LI-AC support the
stacking function.
l Among S5700S-P-LI series switches, only the S5700S-28P-PWR-LI-AC supports the stacking
function.
l Since V200R011C10, S5700LI and S5700S-LI series switches can set up stacks in service port
stacking mode using dedicated stack cables.
l The number of stacked switches is recommended to be 2 to 5.
ur later
SF versions)
P
op
tic
al
po
rts
l 0.5 m and
1.5 m SFP
+
dedicated
stack
cable
(supporte
d in
V200R01
1C10 and
later
versions)
V200R01
1C10 and
later
versions)
l For the cables that can be used as stack cables, see "Cables" in the Hardware Description. Using
other cables may cause exceptions in the stack system, for example, frequent error-down events on
stack ports.
l Combo ports cannot function as stack member ports.
l For stacking-supported service ports, some of them can be used as stack ports, whereas some can be
used as ordinary service ports.
l Since V200R011C10, S5720HI series switches can set up stacks in service port stacking mode using
dedicated stack cables.
l The number of S5700HIs or S5710HIs used to set up a stack is recommended to be 2 to 5, and the
number of S5720HIs used to set up a stack is recommended to be 2 to 9.
Ports on a 4- l 1 m, 3 m, 10Gb
port 10GE and 5 m it/s
interface card SFP+
passive
high-
speed
copper
cables
l 10 m SFP
+ active
high-
speed
copper
cables
l 3 m and
10 m
AOC
cables
l 10GE
SFP+
optical
module
and
optical
fiber
l For the cables that can be used as stack cables, see "Cables" in the Hardware Description. Using other
cables may cause exceptions in the stack system, for example, frequent error-down events on stack ports.
l The number of stacked switches is recommended to be 2 to 5.
l For the cables that can be used as stack cables, see "Cables" in the Hardware Description. Using other
cables may cause exceptions in the stack system, for example, frequent error-down events on stack ports.
l Combo ports cannot function as stack member ports.
l For stacking-supported service ports, some of them can be used as stack ports, whereas some can be used
as ordinary service ports.
l Since V200R011C10, S5710-X-LI series switches can set up stacks in service port stacking mode using
dedicated stack cables.
l The number of stacked switches is recommended to be 2 to 5.
d in
V200R01
1C10 and
later
versions)
l For the cables that can be used as stack cables, see "Cables" in the Hardware Description. Using other
cables may cause exceptions in the stack system, for example, frequent error-down events on stack ports.
l Combo ports cannot function as stack member ports.
l For stacking-supported service ports, some of them can be used as stack ports, whereas some can be used
as ordinary service ports.
l The number of stacked switches is recommended to be 2 to 5.
l For the cables that can be used as stack cables, see "Cables" in the Hardware Description. Using
other cables may cause exceptions in the stack system, for example, frequent error-down events on
stack ports.
l S5720EI series switches are classified into S5720-C-EI, S5720-PC-EI, S5720-X-EI, and S5720-P-EI
based on their uplink interface types.
l In V200R009C00 and later versions, if MPLS-incapable S5720EIs exist in a stack, this stack cannot
have MPLS enabled. If member devices in a stack are running MPLS services, adding MPLS-
incapable S5720EIs to the stack is not allowed. To check whether S5720EIs support MPLS, run the
display device capability command.
l The number of stacked switches is recommended to be 2 to 9.
S5720- Two ports on each l 1 m, 3 m, 5 m Any models of the S5720EI series can
C-EI ES5D21VST000 QSFP+ set up a stack.
and stack card passive high-
S5720- speed cables
PC-EI l QSFP+ optical
module (only
QSFP-40G-
SR4 and
QSFP-40G-
iSR4
supported) and
matching
optical fiber
l For the cables that can be used as stack cables, see "Cables" in the Hardware Description. Using
other cables may cause exceptions in the stack system, for example, frequent error-down events on
stack ports.
l Combo ports cannot function as stack member ports.
l For stacking-supported service ports, some of them can be used as stack ports, whereas some can be
used as ordinary service ports.
l In V200R009C00 and later versions, if MPLS-incapable S5720EIs exist in a stack, this stack cannot
have MPLS enabled. If member devices in a stack are running MPLS services, adding MPLS-
incapable S5720EIs to the stack is not allowed. To check whether S5720EIs support MPLS, run the
display device capability command.
l The number of stacked switches is recommended to be 2 to 9.
Electrical Category 6A 10
ports on a 2- or higher Gbit/s
port 10GE network
electrical cables (If
interface card Category 6
cables are
used, ensure
that the
cables meet
requirements
of TSB-155.)
l For the cables that can be used as stack cables, see "Cables" in the Hardware Description. Using
other cables may cause exceptions in the stack system, for example, frequent error-down events on
stack ports.
l Combo ports cannot function as stack member ports.
l For stacking-supported service ports, some of them can be used as stack ports, whereas some can be
used as ordinary service ports.
l After a UCL group is created using the ucl-group command on S5720LI&S5720S-LI switches, a
stack cannot be set up.
l S5720LI series switches are classified into S5720-P-LI, S5720-TP-LI, and S5720-X-LI based on
their uplink interface types.
l S5720S-LI series switches are classified into S5720S-P-LI, S5720S-TP-LI, and S5720S-X-LI based
on their uplink interface types.
l S5720-28X-LI-24S-AC, S5720-28X-LI-24S-DC, and S5720S-28X-LI-24S-AC do not support
electrical ports. Therefore, they cannot set up stacks using electrical ports.
l MultiGE port on the S5720-28X-PWH-LI-AC cannot be configured as a stack member port.
l In a stack set up using electrical interfaces, when adding a new member switch or replacing a faulty
member switch, to prevent a packet loop caused by a failure to add the new member in case of
insufficient ACL resources, you are advised to configure the service ports used in both the new
switch and the stack as stack ports before connecting the switch to the stack using cables.
l Since V200R011C10, S5720LI and S5720S-LI series switches can set up stacks in service port
stacking mode using dedicated stack cables.
l The number of stacked switches is recommended to be 2 to 9.
Electrical Category 5 or 1
ports on the higher Gbit/s
front panel network
cable
Electrical Category 5 or 1
ports on the higher Gbit/s
front panel network
cable
Electrical Category 5 or 1
ports on the higher Gbit/s
front panel network
cable
Electrical Category 5 or 1
ports on the higher Gbit/s
front panel network
cable
l For the cables that can be used as stack cables, see "Cables" in the Hardware Description. Using
other cables may cause exceptions in the stack system, for example, frequent error-down events on
stack ports.
l Combo ports cannot function as stack member ports.
l For stacking-supported service ports, some of them can be used as stack ports, whereas some can be
used as ordinary service ports.
l After a UCL group is created using the ucl-group command on S5720SI and S5720S-SI switches, a
stack cannot be set up.
l S5720SI series switches are classified into S5720-P-SI and S5720-X-SI based on their uplink
interface types.
l S5720S-SI series switches are classified into S5720S-P-SI and S5720S-X-SI based on their uplink
interface types.
l MultiGE ports of an S5720-14X-PWH-SI-AC cannot be used as stack member ports.
l In V200R009 and earlier versions, S5720-14X-PWH-SI-AC cannot set up a stack with S5720S-SI or
other S5720SI models. The S5720S-SI can set up a stack with all S5720SI models except
S5720-14X-PWH-SI-AC. In V200R010 and later versions, any models of S5720SI and S5720S-SI
can set up a stack. When different models need to set up a stack, they must use stack ports of the
same type. That is, they either all use electrical ports on front panels or all use SFP+ or SFP optical
ports.
l Since V200R010, two SFP+ optical ports of an S5720-14X-PWH-SI-AC can be used as stack
member ports. If SFP+ optical ports of an S5720-14X-PWH-SI-AC have been configured as stack
member ports but the switch needs to be downgraded to V200R009 or an earlier version, delete the
stack member port configuration of SFP+ optical ports before the downgrade. If this configuration is
not deleted before the downgrade, a stack that cannot run normally will be set up after the
downgrade. You need to disable the stacking function to delete the stack member port configuration
of SFP+ optical ports.
l In a stack set up using electrical interfaces, when adding a new member switch or replacing a faulty
member switch, to prevent a packet loop caused by a failure to add the new member in case of
insufficient ACL resources, you are advised to configure the service ports used in both the new
switch and the stack as stack ports before connecting the switch to the stack using cables.
l Since V200R011C10, S5720SI and S5720S-SI series switches can set up stacks in service port
stacking mode using dedicated stack cables.
l The number of stacked switches is recommended to be 2 to 9.
S572 Four SFP l 1 m and 3 10 l When SFP optical ports are used as
0-P- optical ports m SFP+ Gbit/s stack ports, a switch supports a
SI passive maximum of two logical stack ports,
and high- and each logical stack port contains
S572 speed one or two stack member ports. Each
0S-P- copper switch can use a maximum of four
SI cables service ports as stack member ports.
l 5 m SFP+ When a logical stack port contains
passive two stack member ports, the logical
high- stack port can contain only stack
speed member ports 1 and 2 or stack
copper member ports 3 and 4.
cables l When electrical ports on the front
(supporte panel are used as stack ports, a switch
d in supports a maximum of two logical
V200R00 stack ports, and each logical stack
9 and port supports at least one stack
later member port and at most eight stack
versions) member ports. A switch supports a
l 10 m SFP maximum of 16 stack member
+ active ports.On a switch with 48 electrical
high- ports, the first and last 24 ports must
speed be added to two separate logical stack
copper ports to ensure that a stack can be set
cables up successfully.
l 3 m and l Since V200R010, if stack member
10 m ports are connected using 1 m or 3 m
AOC SFP+ passive high-speed copper
cables cables, their working speed can be
increased to 12 Gbit/s using the stack
l 10GE port speed command. After their
SFP+ working speed is increased to 12
optical Gbit/s, switches using these ports
module cannot set up a stack with switches
and using ports with the working speed
optical 10 Gbit/s.
fiber
l 0.5 m and
1.5 m SFP
+
dedicated
stack
cable
(supporte
d in
V200R01
1C10 and
later
versions)
Electrical Category 5 or 1
ports on the higher Gbit/s
front panel network
cable
S572 Four SFP+ l 1 m and 3 10 l When SFP+ optical ports are used as
0-X- optical ports m SFP+ Gbit/s stack ports, a switch supports a
SI passive maximum of two logical stack ports,
and high- and each logical stack port contains
S572 speed one or two stack member ports. Each
0S-X- copper switch can use a maximum of four
SI cables service ports as stack member ports.
l 5 m SFP+ When a logical stack port contains
passive two stack member ports, the logical
high- stack port can contain only stack
speed member ports 1 and 2 or stack
copper member ports 3 and 4.
cables l When electrical ports or downlink
(supporte SFP optical ports on the front panel
d in are used as stack ports, a switch
V200R00 supports a maximum of two logical
9 and stack ports, and each logical stack
later port supports at least one stack
versions) member port and at most eight stack
member ports. A switch supports a
l 10 m SFP
maximum of 16 stack member
+ active
ports.On a switch with 48 electrical
high-
ports, the first and last 24 ports must
speed
be added to two separate logical stack
copper
ports to ensure that a stack can be set
cables
up successfully.
l 3 m and
l Since V200R010, several new
10 m
S5720-X-SI models(S5720-28X-
AOC
SI-24S-AC, S5720-28X-SI-24S-DC,
cables
and S5721-28X-SI-24S-AC)
l 10GE providing downlink SFP optical ports
SFP+ are available. These switches allow
optical using downlink SFP optical ports on
module the front panel as stack member
and ports. If one member switch uses
optical downlink SFP optical ports as stack
fiber member ports, all the other member
l 0.5 m and switches must also use downlink SFP
1.5 m SFP optical ports as stack member ports.
+ l Since V200R010, if stack member
dedicated ports are connected using 1 m or 3 m
stack SFP+ passive high-speed copper
cable cables, their working speed can be
(supporte increased to 12 Gbit/s using the stack
d in port speed command. After their
V200R01
Downlink GE SFP 1
SFP optical optical Gbit/s
ports on the module and
front panel optical fiber
l For the cables that can be used as stack cables, see "Cables" in the Hardware Description. Using
other cables may cause exceptions in the stack system, for example, frequent error-down events on
stack ports.
l Combo ports cannot function as stack member ports.
l For stacking-supported service ports, some of them can be used as stack ports, whereas some can be
used as ordinary service ports.
l S5730SI series switches of all versions can set up stacks in service port stacking mode using
dedicated stack cables.
l The number of stacked switches is recommended to be 2 to 9.
l For the cables that can be used as stack cables, see "Cables" in the Hardware Description. Using
other cables may cause exceptions in the stack system, for example, frequent error-down events on
stack ports.
l Combo ports cannot function as stack member ports.
l For stacking-supported service ports, some of them can be used as stack ports, whereas some can be
used as ordinary service ports.
l S5730S-EI series switches of all versions can set up stacks in service port stacking mode using
dedicated stack cables.
l The number of stacked switches is recommended to be 2 to 9.
l For the cables that can be used as stack cables, see "Cables" in the Hardware Description. Using
other cables may cause exceptions in the stack system, for example, frequent error-down events on
stack ports.
l Combo ports cannot function as stack member ports.
l The number of stacked switches is recommended to be 2 to 5.
l 10GE
SFP+
optical
module
and
optical
fiber
l For the cables that can be used as stack cables, see "Cables" in the Hardware Description. Using
other cables may cause exceptions in the stack system, for example, frequent error-down events on
stack ports.
l Combo ports cannot function as stack member ports.
l Since V200R011C10, S6720EI and S6720S-EI series switches can set up stacks in service port
stacking mode using dedicated stack cables.
l The number of stacked switches is recommended to be 2 to 9.
40GE ports l 1 m, 3 m, 40
on the front and 5 m Gbit/s
panel QSFP+
passive
high-
speed
copper
cables
l QSFP+
optical
module
(QSFP-40
G-SR-BD
not
supported
) and
optical
fiber
l 10 m
QSFP+
AOC
cable
l For the cables that can be used as stack cables, see "Cables" in the Hardware Description. Using
other cables may cause exceptions in the stack system, for example, frequent error-down events on
stack ports.
l Combo ports cannot function as stack member ports.
l For stacking-supported service ports, some of them can be used as stack ports, whereas some can be
used as ordinary service ports.
l Since V200R011C10, S6720SI and S6720S-SI series switches can set up stacks in service port
stacking mode using dedicated stack cables.
l The number of stacked switches is recommended to be 2 to 9.
Electrical Category 6A 10
ports on the or higher Gbit/s
front panel network
cables (If
Category 6
cables are
used, ensure
that the
cables meet
requirements
of TSB-155.)
l For the cables that can be used as stack cables, see "Cables" in the Hardware Description. Using
other cables may cause exceptions in the stack system, for example, frequent error-down events on
stack ports.
l Combo ports cannot function as stack member ports.
l For stacking-supported service ports, some of them can be used as stack ports, whereas some can be
used as ordinary service ports.
l Since V200R011C10, S6720LI and S6720S-LI series switches can set up stacks in service port
stacking mode using dedicated stack cables.
l The number of stacked switches is recommended to be 2 to 9.
Stack ID 0
In Figure 9-29, the three steps in the dashed box are included in 9.9.2 Performing Software
Configurations. Configuring a stack ID and stack priority are optional but are recommended to
facilitate management.
Stack cards do not need to be installed when you use fixed stack ports to establish a stack.
Enable stacking
Connect stack
cables and power on
member switches
Configure stack IDs
for member switches
Check whether
the stack has been
set up successfully
Configure stack
priorities for member
switches
End
Optional
Mandatory
Precautions
Card Storage and Transfer
l Handle a stack card carefully when it is outside the chassis. Wear an ESD wrist strap or
ESD gloves when touching the stack card. Place the stack card on a horizontal plane,
with the electronic components facing upward. Do not place any objects on or near the
card.
l Keep a stack card from heavy humidity or direct sunlight. Ensure that the card is stored
in an environment that will not cause damages to the card.
l To move multiple stack cards, move one card each time. Do not stack the cards together
when moving them.
Card installation and removal
l Wear an ESD wrist strap or ESD gloves when installing or removing a stack card. Do not
touch the printed circuit board (PCB) on the card.
l Gently, slowly pull or push the stack card along the guide rails. Avoid short-circuit
conditions on the card caused by collision with metal materials or improper use of tools.
Installation Procedure
Perform the following steps to install a stack card:
1. Wear an ESD wrist strap and connect the other end of the strap to the ESD jack on the
cabinet.
2. Power off the switch and remove the filler panel from the rear card slot.
3. Install the stack card in the rear card slot, close the ejector levers on the stack card, and
fasten the captive screws. See Figure 9-30.
4. Power on the switch.
Context
After you install a stack card and power on the switch, you can configure related software on
the switch. Available configuration options include configuring a stack ID, and configuring a
stack priority. By default, the stacking function is enabled on a switch, and you are advised to
configure a stack ID and a stack priority to facilitate management. You can configure stack
software on a switch that uses stack card stacking with stack ports fixed on the rear panel after
the switch is powered on, without having to install stack cards on the switch.
Stack IDs identify member switches in a stack. Each stack member switch has a unique stack
ID. If a new member switch has the same stack ID as an existing member switch, the master
switch assigns another stack ID to the new member switch. You can also configure a stack ID
for this member switch. To facilitate management on switches, you are advised to plan and
configure stack IDs for member switches before setting up a stack. A member switch's stack
ID affects many configurations on the switch, for example, port numbers (physical and logical
ports) and port configurations. Pay attention to the following points when configuring stack
IDs for member switches:
l If you do not restart a switch after changing its stack ID, the switch still uses the original
stack ID, and all physical resources are identified by the original stack ID.
l If you save the current configuration and restart the switch after changing its stack ID,
the new stack ID takes effect and all the physical resources are identified by the new
stack ID. In the configuration file, only the global stack configuration and stack priority
of the switch are still in effect. All the other configurations related to the old stack ID
(such as port configuration) become invalid and must be reconfigured.
The stack priority determines the role of member switches during role election. A larger value
indicates a higher priority and higher probability that the member switch will be elected as the
master switch.
Procedure
Step 1 Run the system-view command to enter the system view.
Step 2 (Optional) Run the stack slot slot-id renumber new-slot-id command to configure a stack ID
for the switch.
Step 3 (Optional) Run the stack slot slot-id priority priority command to configure a stack priority
for the switch.
Step 4 (Optional) Run the display stack configuration [ slot slot-id ] command to verify the stack
configuration on a member switch.
----End
The stack configuration is saved automatically. If you want to save other configurations, run the save
command to save them before you power off the switches.
Procedure
Step 1 (Optional) Run the save command to save the configuration.
----End
Precautions
In stack card connection mode, member switches are connected through two ports that reside
on stack cards or are integrated into member switches using dedicated QSFP+ cables or QSFP
+ optical modules and fibers. Figure 9-31 shows a QSFP+ cable.
NOTE
l Wear an ESD wrist strap or ESD gloves when installing PCIe cables.
l Prevent the cables from being twisted.
Installation Procedure
Two stack topologies are available: chain topology and ring topology, as shown in Figure
9-32 and Figure 9-33. For comparison of the two topologies, see 9.3.2 Stack Setup.
NOTE
To ensure that a stack can be set up successfully, you are advised to perform operations in the
following sequence (in a chain topology with three switches A, B, and C):
1. Wear an ESD wrist strap and connect the other end of the strap to the ESD jack on the
cabinet.
NOTE
To specify a member switch as the master switch, power on that switch first. Power on the other member
switches after that switch has started. In this example, SwitchA becomes the master switch after the
stack is set up.
After you complete stack configurations, review key indicators to confirm the stack has been
established successfully. If the stack has been successfully established, log in to the stack and
run commands to confirm that the stack topology matches the hardware connections. If a
stack has not been established, log in to the stack (multiple independent switches) and run
commands to locate and resolve the issues.
9.9.5.1 Reviewing Key Indicators to Confirm that a Stack Has Been Established
Context
After you complete configurations for establishing a stack, you can review indicators on the
switch panels to check whether the stack has been established successfully.
Procedure
1. Press the mode switching button on any member switch to set the mode status indicator
to work in STACK mode. A switch has a STCK indicator, which indicates the stack
mode status. After you press the MODE button, the indicator is steady green or blinking
and off in 45 seconds, indicating that the switch has entered the STACK mode.
– If the mode status indicators on all the member switches change to the STACK
mode, the stack has been established successfully. You can determine the stack IDs
of the master switch and member switches by viewing the mode status indicator and
service port indicators.
– If the mode status indicators on some member switches do not change to the
STACK mode, the stack has not been established.
2. Determine the stack IDs of the master switch and member switches through the status of
service port indicators. Table 9-25 explains how to interpret service port indicators in
STACK mode.
Context
If a stack fails to be established, you can log in to the stack and run commands to check
whether the stack topology corresponds to the hardware connections or to locate and rectify
faults.
Procedure
1. Log in to the stack.
You can log in to a stack using the following methods:
– Local login: Log in through the console interface of any member switch.
– Remote login: Log in through the management interface or another Layer 3
interface of any member switch. You can remotely log in to the stack using Telnet,
STelnet, web, or SNMP if your operation terminal has a reachable route to the
stack.
NOTE
l After switches with management interfaces form a stack, only the management interface
on one member switch takes effect and becomes the master management interface. After
the stack starts, the management interface on the master switch is used as the master
management interface. If the management interface on the master switch is faulty or
unavailable, the management interface on another member switch is used as the master
management interface. If you connect your PC directly to a non-master management
interface, you cannot log in to the stack.
l After a stack is set up, all the member switches use the configuration file of the master
switch.
l A stack externally acts as a single switch and uses one management IP address. To
remotely log in to a stack, you need to log in to the IP address of the master switch.
l All stack member switches are able to forward traffic.
2. Check whether the stack has been established successfully.
Run the display device command to check whether the number of member switches in
the stack is the same as the number of switches in the networking.
<HUAWEI> display device
S5720-32X-EI-AC's Device status:
Slot Sub Type Online Power Register Status Role
------------------------------------------------------------------------------
-
0 - S5720-32X-EI Present PowerOn Registered Normal
Master
1 - S5720-32X-EI Present PowerOn Registered Normal
Standby
– Run the display stack peers command to check whether neighboring information
about the stack is the same as the actual hardware connections.
<HUAWEI> display stack peers
Slot Port1 Peer1 Port2
Peer2
-------------------------------------------------------------------------
--
0 STACK 1 1 STACK 2
None
1 STACK 1 None STACK 2 0
– If they are the same, the stack has been established successfully.
– If they are different, reconnect the member switches.
– If the fault persists, rectify the fault according to Rectifying Common Stack
Faults.
Some member switches may fail to be added to a stack due to the following causes:
l Switches of different models cannot set up a stack.
l A switch has no electronic label or an incorrect electronic label.
l A stack card has no electronic label or an incorrect electronic label.
l Stack ports are incorrectly connected.
l Some stack cables or stack cards are faulty.
l The stack reserved VLAN is occupied.
l The MAC addresses of member switches conflict with each other.
Procedure
– If switches of different models cannot set up a stack or a switch does not support
the stacking function, replace the switch.
– If the switches support stacking, proceed to the next step.
2. Check that a correct electronic label has been loaded to the switch.
Run the display elabel command to view the electronic label.
<HUAWEI> display elabel
/$[System Integration Version]
/$SystemIntegrationVersion=3.0
[Slot_0]
/$[Board Integration Version]
/$BoardIntegrationVersion=3.0
[Main_Board]
/$[ArchivesInfo Version]
/$ArchivesInfoVersion=3.0
[Board Properties]
BoardType=S5720EI
BarCode=21023595860123456789
Item=02359586
Description=S5720-32X-EI-AC(24 Ethernet 10/100/1000 ports,4 100/1000 SFP,4
10 Gig SFP+,AC 110/220V)
Manufactured=2014-05-15
VendorName=Huawei
IssueNumber=00
CLEICode=
BOM=02359586
– If all fields under [Board Properties] are empty, no electronic label is loaded to the
switch. Replace the switch.
– If fields under [Board Properties] are not empty, an electronic label has been loaded
to the switch. Proceed to the next step.
3. Check the electronic label of the stack card. This step is not required on a switch that
uses stack card stacking with stack ports fixed on the rear panel.
Run the display elabel slot slot-id subcard-id command to check the electronic label of
the stack card. slot-id indicates the stack ID and you can obtain it from the output of the
display stack command. subcard-id indicates the subcard ID which is 1 in a stack. For
example, to view the stack card electronic label on a switch with stack ID 0, run the
display elabel slot 0 1 command.
<HUAWEI> display elabel slot 0 1
/$[System Integration Version]
/$SystemIntegrationVersion=3.0
[Slot_0]
/$[Board Integration Version]
/$BoardIntegrationVersion=3.0
[Main_Board]
/$[ArchivesInfo Version]
/$ArchivesInfoVersion=3.0
[Board Properties]
BoardType=S5720EI
BarCode=0302MDW
Item=
Description=Dedicated stack card with 2*QSFP+ interface(Including one PCS of
1M QSFP+ cable ,Used in S5720EI series)
Manufactured=2014-07-07
VendorName=Huawei
IssueNumber=00
CLEICode=
BOM=
– If all fields under [Board Properties] are empty, the stack card is not functioning
properly. Replace the stack card.
– If fields under [Board Properties] are not empty, an electronic label has been loaded
to the switch. Proceed to the next step.
4. Check the stack cables.
Run the display stack port [ brief | slot slot-id ] command to check whether stack
cables are correctly connected.
<HUAWEI> display stack port slot 0
stack-
port0/1:
---------------------------------------------
Current state :
UP
Speed :
21000
369
Discard: 0, Frames:
0
Total Error:
0
CRC: 0, Giants:
0
Jabbers: 0, Fragments:
0
Runts: 0, DropEvents:
0
Alignments: 0, Symbols:
0
Ignoreds:
0
Total Error:
0
Collisions: 0, ExcessiveCollisions:
0
Late Collisions: 0, Deferreds:
0
Buffers Purged: 0
– If the Current state of a stack port on the switch is Up, proceed to the next step.
– If a stack port status is Down, check whether the switch connected to the port has
been powered off or is restarting. If so, check whether the stack port status is Up
after the connected switch restarts. If not, replace the stack cable.
– If the fault persists, proceed to the next step.
5. Check whether the stack reserved VLAN is occupied.
Run the display stack command to check the reserved VLAN of a member switch. By
default, the reserved VLAN is 4093.
Run the display vlan vlan-id command. If the reserved VLAN is not occupied, proceed
to the next step.
<HUAWEI> display vlan 4093
Error: The VLAN does not exist.
– If the reserved VLAN is occupied, run either of the following commands to rectify
the fault:
n Run the stack reserved-vlan vlan-id command in the system view to modify
the reserved VLAN.
n Delete services in the reserved VLAN and run the undo vlan vlan-id
command in the system view to delete the VLAN.
In Figure 9-34, the three steps in the dashed box are included in 9.10.1 Performing Software
Configurations. Configuring a stack ID and stack priority are optional but are recommended to
facilitate management.
Add physical
member ports to
logical stack ports Power off all
member switches
Configure stack
priorities for member Check whether
switches the stack has been
set up successfully
End
Optional
Mandatory
stack ID. In the configuration file, only the global stack configuration and stack priority
of the switch are still in effect. All the other configurations related to the old stack ID
(such as port configuration) become invalid and must be reconfigured.
The stack priority determines the role of member switches during role election. A larger value
indicates a higher priority and higher probability that the member switch will be elected as the
master switch.
Procedure
Step 1 Run the system-view command to enter the system view.
Step 2 Run the interface stack-port member-id/port-id command to create a logical stack port and
enter the logical stack port view.
1. A stack physical member port supports only stack-related functions, and other functions cannot be
configured on the interface. All the commands irrelevant to the stack function are unavailable in the
interface view, and only basic configuration commands such as description (interface view) are
retained.
2. After configuring a service port of a switch as a physical member port, you are advised to save the
configuration if this service port has been referenced by other commands. Otherwise, the commands
that reference this service port may be retained after the switch restarts.
3. To restore a physical member port as a service port, run the shutdown interface { interface-type
interface-number1 [ to interface-type interface-number2 ] } &<1-10> command in the logical stack
port view to shut down the physical member port, and then run the undo port interface { interface-
type interface-number1 [ to interface-type interface-number2 ] } &<1-10> enable command in this
view.
4. On the S6720EI and S6720S-EI, every four of XGE interfaces from the left are added to one group.
For example, XGE interfaces numbered 1 to 4 can be added to one group, but XGE interfaces
numbered 2 to 5 cannot. That is, the number of the last XGE interface in each group must be the
multiple of 4. If you configure any interface in each group as a physical member port, configurations
on the other three interfaces in the group will be lost and the three interfaces cannot be used as
service ports.
Step 5 (Optional) Run the stack slot slot-id renumber new-slot-id command to configure a stack ID
for the switch.
Step 6 (Optional) Run the stack slot slot-id priority priority command to configure a stack priority
for the switch.
Step 7 (Optional) Run the display stack configuration [ slot slot-id ] command to verify the stack
configuration on a member switch.
----End
Context
After you complete the preceding configurations, power off all member switches.
NOTE
The stack configuration is saved automatically. If you want to save other configurations, run the save
command to save them before you power off the switches.
Procedure
Step 1 (Optional) Run the save command to save the configuration.
----End
Precautions
When installing or removing optical fibers, do not look into optical ports or connectors
without eye protection.
Installation Procedure
Stack member switches can form a chain or ring topology through service port connections.
Figure 9-35 and Figure 9-36 shows stack cable connections in a stack that is set up by three
S5700-28X-LI-AC switches. On the member switches, the first two 10GE optical ports are
configured as physical member ports of logical stack port 1, and the last 10GE optical ports
are configured as physical member ports of logical stack port 2.
NOTE
Figure 9-35 Ring topology of a stack set up through service port connections
Logical
stack port 2
Figure 9-36 Chain topology of a stack set up through service port connections
Logical
stack port 2
Logical
stack port 2
To ensure that a stack can be set up successfully, you are advised to perform operations in the
following sequence (in a chain topology with three switches A, B, and C):
1. Wear an ESD wrist strap and connect the other end of the strap to the ESD jack on the
cabinet.
2. Use stack cables to connect SwitchA and SwitchB.
3. Power on SwitchA and then SwitchB.
4. Check whether SwitchA and SwitchB have set up a stack successfully. (See 9.10.4
Checking Whether a Stack Is Established Through Service Port Connections.)
5. Use stack cables to connect SwitchC and SwitchB, and then power on SwitchC.
6. Check whether SwitchA, SwitchB, and SwitchC have set up a stack successfully. (See
9.10.4 Checking Whether a Stack Is Established Through Service Port
Connections.)
NOTE
To specify a member switch as the master switch, power on that switch first. Power on the other member
switches after that switch has started. In this example, SwitchA becomes the master switch after the
stack is set up.
has not been established, log in to the stack (multiple independent switches) and run
commands to locate and resolve the issues.
9.10.4.1 Reviewing Key Indicators to Confirm that a Stack Has Been Established
Context
After the stack powered on, review key indicators on the switch panels to confirm the stack
has been established successfully.
Procedure
l Press the mode switching button on any member switch to set the mode status indicator
to work in STACK mode. A switch has a STCK indicator, which indicates the stack
mode status. After you press the MODE button, the indicator is steady green or blinking
and off in 45 seconds, indicating that the switch has entered the STACK mode.
– If the mode status indicators on all the member switches change to the STACK
mode, the stack has been established successfully. You can determine the stack IDs
of the master switch and member switches by viewing the mode status indicator and
service port indicators.
– If the mode status indicators on some member switches do not change to the
STACK mode, the stack has not been established.
l Determine the stack IDs of the master switch and member switches through the status of
service port indicators. Table 9-26 explains how to interpret service port indicators in
STACK mode.
Context
If a stack fails to be established, you can log in to the stack and run commands to check
whether the stack topology corresponds to the hardware connections or to locate and rectify
faults.
Procedure
1. Log in to the stack.
You can log in to a stack using the following methods:
– Local login: Log in through the console interface of any member switch.
– Remote login: Log in through the management interface or another Layer 3
interface of any member switch. You can remotely log in to the stack using Telnet,
STelnet, web, or SNMP if your operation terminal has a reachable route to the
stack.
NOTE
l After switches with management interfaces form a stack, only the management interface
on one member switch takes effect and becomes the master management interface. After
the stack starts, the management interface on the master switch is used as the master
management interface. If the management interface on the master switch is faulty or
unavailable, the management interface on another member switch is used as the master
management interface. If you connect your PC directly to a non-master management
interface, you cannot log in to the stack.
l After a stack is set up, all the member switches use the configuration file of the master
switch.
l A stack externally acts as a single switch and uses one management IP address. To
remotely log in to a stack, you need to log in to the IP address of the master switch.
l All stack member switches are able to forward traffic.
2. Check whether the stack has been established successfully.
Run the display device command to check whether the number of member switches in
the stack is the same as the number of switches in the networking.
<HUAWEI> display device
S5700-28P-LI-AC's Device status:
Slot Sub Type Online Power Register Status Role
------------------------------------------------------------------------------
-
0 - S5700-28P-LI Present PowerOn Registered Normal
Master
1 - S5700-28P-LI Present PowerOn Registered Normal
Standby
– Run the display stack peers command to check whether neighboring information
about the stack is the same as the actual hardware connections.
<HUAWEI> display stack peers
Slot Port1 Peer1 Port2
Peer2
-------------------------------------------------------------------------
--
0 STACK 1 1 STACK 2
None
1 STACK 1 None STACK 2 0
– Run the display stack port command to view information about physical member
ports added to a logical stack port.
<HUAWEI> display stack port
*down : administratively
down
– Run the display stack channel all command to check the stack link connections
and status.
<HUAWEI> display stack channel all
! : Port have received packets with CRC
error.
P-Port: Physical
port
– If they are the same, the stack has been established successfully.
– If they are different, reconnect the member switches.
– If the fault persists, rectify the fault according to Rectifying Common Stack
Faults.
– If switches of different models cannot set up a stack or a switch does not support
the stacking function, replace the switch.
– If the switches support stacking, proceed to the next step.
2. Check that a correct electronic label has been loaded to the switch.
Run the display elabel command to view the electronic label.
<HUAWEI> display elabel
/$[System Integration Version]
/$SystemIntegrationVersion=3.0
[Slot_0]
/$[Board Integration Version]
/$BoardIntegrationVersion=3.0
[Main_Board]
/$[ArchivesInfo Version]
/$ArchivesInfoVersion=3.0
[Board Properties]
BoardType=CX22EMGEB
BarCode=21023518320123456789
Item=02351832
Description=S5700-28P-LI,LS5ZC48CM,S5700-28P-LI Mainframe
Manufactured=2009-02-05
VendorName=Huawei
IssueNumber=
CLEICode=
BOM=
– If all fields under [Board Properties] are empty, no electronic label has been loaded
to the switch. Replace the switch.
– If fields under [Board Properties] are not empty, an electronic label has been loaded
to the switch. Proceed to the next step.
3. Check the stack cables.
Run the display stack port command to check whether stack cables are correctly
connected.
<HUAWEI> display stack port
*down : administratively down
Logic Port Phy Port Online Status
----------------------------------------------------------------------------
stack-port0/1 GigabitEthernet0/0/1 present up
stack-port1/2 GigabitEthernet1/0/1 present up
– If the physical member ports are in the Up state, proceed to the next step.
– If the status of a physical member port is in the Down state, check whether the
switch connected to the port has been powered off or is restarting. If so, check
whether the port status is in the Up state after the connected switch restarts. If not,
replace the stack cable.
– If the reserved VLAN 4093 is occupied, run either of the following commands to
rectify the fault:
n Run the stack reserved-vlan vlan-id command in the system view to modify
the reserved VLAN.
n Delete services in the reserved VLAN and run the undo vlan vlan-id
command in the system view to delete the VLAN.
– If the fault persists, proceed to the next step.
5. Check whether MAC addresses of member switches conflict with each other.
Run the display stack command to check whether the MAC address of the switch to be
added to the stack conflicts with the MAC address of an existing member switch. If not,
proceed to the next step.
<HUAWEI> display stack
Stack mode: Service-port
Stack topology type: Link
Stack system MAC: 0018-82b1-6eb4
MAC switch delay time: 2 min
Stack reserved vlan: 4093
Slot of the active management port: --
Slot Role Mac address Priority Device type
-------------------------------------------------------------
0 Master 0018-82b1-6eb4 200 S5700-28P-LI-AC
1 Standby 0018-82b1-6eba 150 S5700-28P-LI-AC
Installation Preparation
l Components: dedicated stack cables
l Tools: ESD wrist strap or ESD gloves
Precautions
l Wear an ESD wrist strap or ESD gloves during the installation.
l Ensure that the stack cables are not tangled with other cables.
l To remove a cable, gently push the connector of the cable and then pull out the cable by
the pull ring.
Cable Connections
l Logical stack port 1 on one switch must be connected to logical stack port 2 on another
switch. Logical stack port numbers have been preconfigured for physical ports as shown
in Table 9-27. To view the port numbers, run the display stack port auto-cable-info
command.
Two uplink optical ports Uplink optical port 1 Uplink optical port 2
(non-combo ports) on the
front panel
Four uplink optical ports First two uplink optical Last two uplink optical
on the front panel ports ports
Eight uplink optical ports Uplink optical ports 1, 2, Uplink optical ports 3, 4,
on the front panel 5, and 6 7, and 8
Last 16 optical ports Optical ports 9, 10, 13, 14, Optical ports 11, 12, 15,
among all 24 optical ports 17, 18, 21, and 22 16, 19, 20, 23, and 24
on the front panel
Last 16 optical ports Optical ports 17, 18, 21, Optical ports 19, 20, 23,
among all 32 optical ports 22, 25, 26, 29, and 30 24, 27, 28, 31, and 32
on the front panel
Last 16 optical ports Optical ports 33, 34, 37, Optical ports 35, 36, 39,
among all 48 optical ports 38, 41, 42, 45, and 46 40, 43, 44, 47, and 48
on the front panel
l Stack member ports of a logical stack port can be connected at any sequence. Each
logical stack port can have one or more physical ports connected.
l If logical stack port numbers have been configured manually, these port numbers will not
change after stack cables are connected. The preconfigured logical stack port numbers do
not take effect. Connect stack cables according to the configured port numbers.
l In unconfigured deployment scenarios, you are advised to connect member switches
based on the following rules:
– Connect the switches in sequence from top to bottom, as shown in Figure 9-37.
– Ensure that all logical stack ports of the top switch are connected to the master ends
of cables, all logical stack ports of the bottom switch are connected to the slave
ends of cables, and two logical stack ports of the intermediate switch are connected
to the master and slave ends respectively.
– After the switches have been connected using dedicated stack cables, they
automatically set up a stack and their stack IDs as well as stack roles are
automatically assigned.
– If the switches are not connected in a ring topology, you only need to ensure that
logical stack port 1 of the local switch is connected to logical stack port 2 of the
remote switch. In this situation, these switches can set up a stack, but their master
and standby roles and stack IDs are randomly generated.
l To perform stack capacity expansion on a switch that already has the configuration, pay
attention to the following:
– Ports with dedicated stack cables installed cannot have any service configuration.
– You only need to ensure that logical stack port 1 of one switch is connected to
logical stack port 2 of another switch, without having to consider the master/slave
ends of dedicated stack cables.
– The existing stack ID of the switch is used when no stack ID conflict occurs. If such
a conflict occurs, a new stack ID is automatically assigned to the switch.
Installation Procedure
Stack member switches can set up a stack in a chain or ring topology through service port
connections. Figure 9-37 and Figure 9-38 show stack cable connections in a stack that is set
up by three S5720-28X-LI-AC switches in ring and chain topologies respectively. On these
member switches, the first two 10GE optical ports are preconfigured as stack member ports of
logical stack port 1, and the last two 10GE optical ports are preconfigured as stack member
ports of logical stack port 2. Here, each logical stack port has two physical member ports
connected and can also have one physical member port connected. To ensure reliability, it is
recommended that a logical stack port have two or more physical member ports connected.
NOTE
l Before using dedicated stack cables to connect member switches, you are advised to power off all
member switches to ensure security.
l When the switch is running and a dedicated stack cable is removed and inserted to a port, this port
changes between a service port and a stack port after 60 seconds to prevent flapping in the event that
this cable is removed and inserted to this port multiple times in a short time. That is, this port
becomes a stack port 60 seconds after a dedicated stack cable is inserted to it or it becomes a service
port 60 seconds after this cable is removed from it. If the stack configuration has been saved before
the dedicated stack cable is removed from a stack port, this stack port will not automatically become
a service port.
l If ports cannot automatically become stack ports after they have dedicated stack cables connected,
an alarm 1.3.6.1.4.1.2011.5.25.183.1.22.64 hwStackAutoConfigFailed is generated. Solve this
problem based on the possible causes and procedure described in this alarm.
Figure 9-37 Stack set up in a ring topology using dedicated stack cables
SwitchA
SwitchB
SwitchC
Figure 9-38 Stack set up in a chain topology using dedicated stack cables
SwitchA
SwitchB
SwitchC
l If the switches have been powered off before they have stack cables connected, you are
advised to power on them in sequence. If you want a specific switch to become the
master switch, power on it first. After it starts, power on other switches. In the following
example, three switches, SwitchA, SwitchB, and SwitchC set up a stack in a chain
topology, and SwitchA becomes the master switch after the three switches are powered
on in the following sequence:
a. Power on and start SwitchA and then power on SwitchB.
b. Check whether SwitchA and SwitchB set up a stack successfully according to
9.10.4 Checking Whether a Stack Is Established Through Service Port
Connections.
c. Power on SwitchC.
d. Check whether SwitchA, SwitchB, and SwitchC set up a stack successfully
according to 9.10.4 Checking Whether a Stack Is Established Through Service
Port Connections.
l If the switches are not powered off before they have stack cables connected, their roles
and stack IDs will be randomly assigned.
After the stack powered on, review key indicators to confirm the stack has been established
successfully. If the stack has been successfully established, log in to the stack and run
commands to confirm that the stack topology matches the hardware connections. If a stack
has not been established, log in to the stack (multiple independent switches) and run
commands to locate and resolve the issues.
9.11.2.1 Reviewing Key Indicators to Confirm that a Stack Has Been Established
Context
After the stack powered on, review key indicators on the switch panels to confirm the stack
has been established successfully.
Procedure
l Press the mode switching button on any member switch to set the mode status indicator
to work in STACK mode. A switch has a STCK indicator, which indicates the stack
mode status. After you press the MODE button, the indicator is steady green or blinking
and off in 45 seconds, indicating that the switch has entered the STACK mode.
– If the mode status indicators on all the member switches change to the STACK
mode, the stack has been established successfully. You can determine the stack IDs
of the master switch and member switches by viewing the mode status indicator and
service port indicators.
– If the mode status indicators on some member switches do not change to the
STACK mode, the stack has not been established.
l Determine the stack IDs of the master switch and member switches through the status of
service port indicators. Table 9-28 explains how to interpret service port indicators in
STACK mode.
Context
If a stack fails to be established, you can log in to the stack and run commands to check
whether the stack topology corresponds to the hardware connections or to locate and rectify
faults.
Procedure
1. Log in to the stack.
You can log in to a stack using the following methods:
– Local login: Log in through the console interface of any member switch.
– Remote login: Log in through the management interface or another Layer 3
interface of any member switch. You can remotely log in to the stack using Telnet,
STelnet, web, or SNMP if your operation terminal has a reachable route to the
stack.
NOTE
l After switches with management interfaces form a stack, only the management interface
on one member switch takes effect and becomes the master management interface. After
the stack starts, the management interface on the master switch is used as the master
management interface. If the management interface on the master switch is faulty or
unavailable, the management interface on another member switch is used as the master
management interface. If you connect your PC directly to a non-master management
interface, you cannot log in to the stack.
l After a stack is set up, all the member switches use the configuration file of the master
switch.
l A stack externally acts as a single switch and uses one management IP address. To
remotely log in to a stack, you need to log in to the IP address of the master switch.
l All stack member switches are able to forward traffic.
2. Check whether the stack has been established successfully.
Run the display device command to check whether the number of member switches in
the stack is the same as the number of switches in the networking.
<HUAWEI> display device
S5720-28P-LI-AC's Device status:
Slot Sub Type Online Power Register Status Role
------------------------------------------------------------------------------
-
0 - S5720-28P-LI Present PowerOn Registered Normal
Master
1 - S5720-28P-LI Present PowerOn Registered Normal
Standby
4093
-------------------------------------------------------------
0 Master 0018-82b1-6eb4 100 S5720-28P-LI-AC
1 Standby 0018-82b1-6eba 100 S5720-28P-LI-AC
– Run the display stack peers command to check whether neighboring information
about the stack is the same as the actual hardware connections.
<HUAWEI> display stack peers
Slot Port1 Peer1 Port2
Peer2
-------------------------------------------------------------------------
--
0 STACK 1 1 STACK 2
None
1 STACK 1 None STACK 2 0
– Run the display stack port command to view information about physical member
ports added to a logical stack port.
<HUAWEI> display stack port
*down : administratively
down
– Run the display stack channel all command to check the stack link connections
and status.
<HUAWEI> display stack channel all
! : Port have received packets with CRC
error.
P-Port: Physical
port
– If they are the same, the stack has been established successfully.
– If they are different, reconnect the member switches.
– If the fault persists, rectify the fault according to Rectifying Common Stack
Faults.
Some member switches may fail to be added to a stack for the following reasons:
l Switches of different models cannot set up a stack.
l A switch does not support the stacking function.
l A switch has no electronic label or an incorrect electronic label.
l Stack ports are incorrectly connected.
l Stack cables are faulty.
l The stack reserved VLAN is occupied.
l MAC addresses of stack member switches conflict.
Procedure
– If switches of different models cannot set up a stack or a switch does not support
the stacking function, replace the switch.
– If the switches support stacking, proceed to the next step.
2. Check that a correct electronic label has been loaded to the switch.
Run the display elabel command to view the electronic label.
<HUAWEI> display elabel
/$[System Integration Version]
/$SystemIntegrationVersion=3.0
[Slot_0]
/$[Board Integration Version]
/$BoardIntegrationVersion=3.0
[Main_Board]
/$[ArchivesInfo Version]
/$ArchivesInfoVersion=3.0
[Board Properties]
BoardType=CX22EMGEB
BarCode=21023518320123456789
Item=02351832
Description=S5720-28P-LI,LS5ZC48CM,S5720-28P-LI Mainframe
Manufactured=2009-02-05
VendorName=Huawei
IssueNumber=
CLEICode=
BOM=
– If all fields under [Board Properties] are empty, no electronic label has been loaded
to the switch. Replace the switch.
– If fields under [Board Properties] are not empty, an electronic label has been loaded
to the switch. Proceed to the next step.
3. Check the stack cables.
Run the display stack port command to check whether stack cables are correctly
connected.
<HUAWEI> display stack port
*down : administratively down
Logic Port Phy Port Online Status
----------------------------------------------------------------------------
stack-port0/1 GigabitEthernet0/0/1 present up
stack-port1/2 GigabitEthernet1/0/1 present up
– If the physical member ports are in the Up state, proceed to the next step.
– If the status of a physical member port is in the Down state, check whether the
switch connected to the port has been powered off or is restarting. If so, check
whether the port status is in the Up state after the connected switch restarts. If not,
replace the stack cable.
– If the fault persists, proceed to the next step.
4. Check whether the stack reserved VLAN is occupied.
Run the display stack command to check the reserved VLAN of a member switch. By
default, the reserved VLAN is 4093.
Run the display vlan vlan-id command. If the reserved VLAN is not occupied, proceed
to the next step.
<HUAWEI> display vlan 4093
Error: The VLAN does not exist.
– If the reserved VLAN is occupied, run either of the following commands to rectify
the fault:
n Run the stack reserved-vlan vlan-id command in the system view to modify
the reserved VLAN.
n Delete services in the reserved VLAN and run the undo vlan vlan-id
command in the system view to delete the VLAN.
– If the fault persists, proceed to the next step.
5. Check whether MAC addresses of member switches conflict with each other.
Run the display stack command to check whether the MAC address of the switch to be
added to the stack conflicts with the MAC address of an existing member switch. If not,
proceed to the next step.
<HUAWEI> display stack
Stack mode: Service-port
Stack topology type: Link
Stack system MAC: 0018-82b1-6eb4
MAC switch delay time: 2 min
Stack reserved vlan: 4093
Slot of the active management port: --
Slot Role Mac address Priority Device type
-------------------------------------------------------------
0 Master 0018-82b1-6eb4 200 S5720-28P-LI-AC
1 Standby 0018-82b1-6eba 150 S5720-28P-LI-AC
Context
After a stack has been set up using dedicated stack cables, the stack configuration
automatically generated will not be automatically saved to the flash memory. To ensure that
the stack configuration still takes effect when these cables are removed or other cables are
connected, save the automatically generated stack configuration to the flash memory.
NOTE
No stack configurations cannot be manually modified before the automatically generated stack configuration
is saved to the flash memory. The system displays error information: "Error: The operation failed because
there was unsaved stack configuration. Please save the stack configuration first."
Procedure
Step 1 Run the system-view command to enter the system view.
Step 2 (Optional) Run the display stack configuration command to view the stack configuration.
If there is configuration starting with a number sign (#) in the command output, this
configuration is not saved to the flash memory. In this situation, no stack configurations can
be modified.
Step 3 You can use the following methods to save the stack configuration automatically generated
for dedicated cable stacking to the flash memory.
l Run the save stack configuration or save command to save the automatically generated
stack configuration to the flash memory.
l If the switch has a configuration file, the stack configuration automatically generated for
dedicated cable stacking is saved to the flash memory after the switch is restarted with
power on.
l If ASs in an SVF system are correctly connected using dedicated stack cables, the stack
configuration automatically generated for dedicated cable stacking is saved to the flash
memory.
If dedicated stack cables are removed and the stack configuration automatically
generated for dedicated cable stacking is not saved to the flash memory, this stack
configuration is cleared.
Step 4 Run the display stack configuration command to check whether all stack configurations are
saved to the flash memory.
----End
Context
MAD detects multiple master switches after a stack splits.
NOTE
You are advised to configure MAD to minimize the impact of a stack split on services.
Configuration Procedure
Table 9-30 describes the MAD configuration procedure.
1 Config You can configure Configure MAD in direct The direct and relay
uring MAD in direct mode when member modes cannot be
MAD mode to perform switches in a stack have configured simultaneously
in MAD through idle ports. Use common in a stack.
Direct dedicated direct cables to connect these By default, MAD is not
Mode links between ports and use the ports for configured on a device.
member switches. MAD only.
4 (Optio You can restore a When MAD detects a Perform this configuration
nal) shutdown port to stack split, multiple stacks when the stack in the
Restori the Up state to compete with each other. Detect state fails. When
ng allow the stack in The stack that wins the the stack in the Detect
Shutdo the Recovery state competition remains in the state is working normally,
wn to work when the Detect state (normal do not perform this
Ports stack in the Detect working state) and the configuration.
to the state fails. stacks that fail the
Up competition enter the
State Recovery state (disabled
state). In the Recovery
state, all the service ports
except reserved ports on
member switches are shut
down, so the stacks do not
forward service packets.
You can restore shutdown
ports to the Up state so
that the stacks in
Recovery state can work
again. For example, if the
stack in Detect state fails
or is removed from the
network before the stack
fault is rectified, restore
the shutdown ports in a
stack in the Recovery state
to the Up state, so that the
stack in the Recovery state
can take over services
from the original active
stack. This minimizes the
impact of a stack fault on
services.
Guidelines
l MAD in direct mode can only be configured on Layer 2 Ethernet ports in Up state.
l You can configure a maximum of eight direct detection links for each member switch in
a stack to improve detection reliability. Ensure that one direct detection link on each
member switch in a stack works normally when multiple master switches coexist.
l MAD is a Huawei proprietary protocol, and Huawei S series switches support the MAD
relay function.
l You can configure the relay mode on a maximum of four Eth-Trunks in a stack to
improve detection reliability. Ensure that one Eth-Trunk in each stack works normally
when multiple master switches coexist.
l When configuring MAD in relay mode, ensure that member switches in a stack use
different MAC addresses. If they use the same MAC address, the proxy device cannot
forward MAD packets.
Procedure
Step 1 Configure MAD in direct or relay mode.
l Configure MAD in direct mode.
a. Run the system-view command to enter the system view.
b. Run the interface interface-type interface-number command to enter the interface
view.
c. Run the mad detect mode direct command to configure MAD in direct mode.
NOTE
– MAD packets are bridge protocol data units (BPDUs). If MAD is performed through dedicated
direct links between member switches and an intermediate device, configure port-based Layer 2
protocol transparent transmission on the intermediate device. For details, see Configuring Interface-
based Layer 2 Protocol Transparent Transmission in "Layer 2 Protocol Transparent Transmission
Configuration" in the S1720, S2700, S5700, and S6720 V200R011C10 Configuration Guide -
Ethernet Switching.
– After MAD in direct mode is configured on an interface, the STP status of the interface becomes
Discarding, affecting the transmission of data packets and some protocol packets. Therefore, do not
configure other services on this interface.
l Configure MAD in relay mode.
– Switch functioning as the relay device
n In the stack
1) Run the system-view command to enter the system view.
2) Run the interface eth-trunk trunk-id command to enter the Eth-Trunk
interface view.
3) Run the mad detect mode relay command to configure MAD in relay
mode on the Eth-Trunk.
n On the specified relay device
1) Run the system-view command to enter the system view.
2) Run the interface eth-trunk trunk-id command to enter the Eth-Trunk
interface view.
3) Run the mad relay command to enable the relay function on the Eth-
Trunk.
– Two stack systems functioning as relay of each other
n In each stack
1) Run the system-view command to enter the system view.
2) Run the mad domain domain-id command to specify the MAD domain
ID for a stack.
By default, the MAD domain ID of a stack is 0.
NOTE
Two stack systems can function as proxy of each other to implement MAD. The
two stack systems must be configured with different MAD domain IDs.
Step 3 (Optional) Run the mad exclude interface { interface-type interface-number1 [ to interface-
type interface-number2 ] } &<1-10> command to configure specified ports in a stack as
reserved ports.
By default, only physical member ports of a stack are reserved ports, and all service ports are
non-reserved ports.
The port used for MAD does not need to be specified as a reserved port. After a stack splits,
the port used for MAD will also be shut down.
Step 4 (Optional) Run the mad restore command to restore shutdown ports to the Up state.
NOTE
Do not run this command when the stack in the Detect state is working normally. Otherwise, there will be
multiple master switches on the network after the stack in the Recovery state is enabled.
If the stack in the Detect state is merged with the stack in the Recovery state, the new stack remains in
Recovery state. You need to run the mad restore command to restore shutdown ports to the Up state.
----End
Context
A stack communicates with other network devices as one device and has a unique MAC
address. This MAC address is the stack MAC address. Generally, the MAC address of the
master switch is used as the stack MAC address.
If the owner of the stack MAC address leaves the stack, switching of the stack MAC address
will cause traffic interruption. You can configure a system MAC address switching delay to
reduce the impact of switching of the stack MAC address on services.
l If the stack MAC address is the MAC address of the master switch and the master switch
fails or leaves the stack, the following may occur:
– When no system MAC address switching delay is configured, the MAC address of
the new master switch is immediately used as the stack MAC address. By default, a
system MAC address switching delay is configured and the switching delay is set to
10 minutes.
– When a system MAC address switching delay is configured, the new master switch
uses its own MAC address as the stack MAC address if the old master switch does
not rejoin the stack within the switching delay; the stack MAC address remains
unchanged if the old master switch rejoins the stack as a slave switch within the
switching delay.
l If the stack MAC address is the MAC address of a slave switch, the MAC address of the
master switch is used as the stack MAC address after the slave switch leaves the stack
and does not rejoin the stack within the switching delay.
Procedure
Step 1 Run the system-view command to enter the system view.
Step 2 Run the stack timer mac-address switch-delay delay-time command to configure a system
MAC address switching delay.
By default, the system MAC address switching delay is 10 minutes.
When the delay time is set to 0, the stack MAC address will not change after the owner of the
stack MAC address leaves the stack. If you run the undo stack timer mac-address switch-
delay command, the stack MAC address will change immediately after the owner of the stack
MAC address leaves the stack.
When the system is configured to switch the system MAC address immediately, the system
uses the MAC address of the new master switch as the system MAC address immediately
when the previous master switch fails or leaves the stack. This may cause protocols such as
LACP and STP to flap, affecting services.
----End
NOTE
Procedure
Step 1 Run the stack led enable [ duration duration-value ] command to enable port indicators to
show stack IDs of member switches.
By default, port indicators do not show stack IDs of member switches.
NOTE
This command can be executed only when the MODE button is not pressed on the switch to change the
indicator mode.
The configuration of this command becomes invalid when port indicators show stack IDs of member
switches for the specified time.
----End
The reserved VLAN of a stack cannot be used to deploy other services. If the reserved VLAN is
occupied, a stack cannot be established successfully. You can run the undo vlan vlan-id command to
delete the occupied VLAN or run the following commands to configure another VLAN as the stack's
reserved VLAN.
Procedure
Step 1 Run the system-view command to enter the system view.
Step 2 Run the stack reserved-vlan vlan-id command to configure a reserved VLAN for a stack.
By default, a stack uses VLAN 4093 as the reserved VLAN.
----End
mode for stack links to prevent traffic congestion by ensuring that outgoing traffic is properly
distributed to multiple stack links. Configure a load balancing mode in the system view or
logical stack port view based on the source IP address, destination IP address, source MAC
address, destination MAC address carried in packets, or a combination of these factors.
l The load balancing mode configured in the system view applies to all logical stack ports.
l The load balancing mode configured in the logical stack port view applies to that port
only.
l A logical stack port gives preference to the load balancing mode configured in the
logical stack port view. If no load balancing mode is configured in the logical stack port
view, the port uses the mode configured in the system view.
You can run the display stack-port command to view the load balancing mode on a logical
stack port.
NOTE
Only the S5720HI, S6720EI, and S6720S-EI series support load balancing for stack links.
On the S5720HI, this command can only be configured in the system view.
Procedure
l Configure a global load balancing mode.
a. Run the system-view command to enter the system view.
b. Run the stack-port load-balance mode { dst-ip | dst-mac | src-dst-ip | src-dst-
mac | src-ip | src-mac } command to configure a global load balancing mode.
By default, the global load balancing mode is ENHANCED.
l Configure a load balancing mode on a logical stack port.
a. Run the system-view command to enter the system view.
b. Run the interface stack-port member-id/port-id command to enter the logical stack
port view.
c. Run the stack-port load-balance mode { dst-ip | dst-mac | src-dst-ip | src-dst-
mac | src-ip | src-mac } command to configure a load balancing mode for the
logical stack port.
By default, the load balancing mode of a logical stack port is ENHANCED.
----End
Only the S5720HI, S6720EI, and S6720S-EI series support load balancing for stack links.
On the S5720HI, this command can only be configured in the system view.
affecting network services. The stack port error-down function can shut down a stack port and
switch traffic to other stack links if the rate of received CRC error packets or the number of
Up/Down transitions on the stack port reaches the specified threshold. This reduces the
impact of a stack port error-down event on network services. Additionally, the system
generates stack port error-down alarms to help in fault location.
When one end is a stack port and the other end is a service port, the system does not set the
stack port to error-down state or generate an error-down alarm even if this stack port
continuously receives CRC error packets.
Procedure
Step 1 Run the stack port { crc | link-flap } trigger error-down command to enable the stack port
error-down function.
By default, the stack port error-down function is enabled.
Step 2 Set stack port error-down parameters.
l Run the stack port { crc | link-flap } trigger { threshold threshold | interval interval }
*command to set the stack port error-down alarm threshold and check interval.
By default, the stack port error-down alarm threshold is 10 times per minute, and the
check interval is 3 minutes. That is, the system will set a stack port to error-down state
and generate an error-down alarm in the following condition: The stack port receives 10
CRC error packets per minute or alternates between Up and Down states 10 times per
minute, and this condition lasts for 3 minute.
l Run the stack port { crc | link-flap } trigger error-down auto-recovery-interval auto-
recovery-interval command to set the stack port error-down alarm clearance interval.
By default, the alarm clearance interval is 0, which means that a stack port error-down
alarm will not be cleared automatically after it is printed.
----End
Procedure
l Run the display stack command to check information about member switches.
l Run the display stack peers command to check information about the neighbors of a
member switch in the stack.
l Run the display stack port [ brief | slot slot-id ] command to check information about
stack ports.
l Run the display stack configuration [ slot slot-id ] command to check the stack
commands that have been executed in the stack.
l Run the display stack channel [ all | slot slot-id ] command to check stack link
connections and status.
----End
Context
After you enable stack traps on the switch, the switch sends trap messages to the network
management system (NMS) when the stack status changes. By default, all stack traps are
enabled. You can use commands to disable all or specified stack traps. Then the switch no
longer sends these traps to the NMS.
To check the status (enabled or disabled) of stack traps, run the display snmp-agent trap
feature-name stack all command.
NOTE
Procedure
l Enable stack traps.
Run the snmp-agent trap enable feature-name stack [ trap-name trap-name ]
command to enable a specified stack trap or all stack traps.
l Disable stack traps.
Run the undo snmp-agent trap enable feature-name stack [ trap-name trap-name ]
command to disable a specified stack trap or all stack traps.
----End
Context
If you want another switch in the stack to be the master switch, perform a master/standby
switchover to change a standby switch to the new master switch.
After the master/standby switchover is complete, the role of each member switch in the stack
changes as shown in Figure 9-39.
SwitchA SwitchA
(Master) (Slave)
Switchover
Stack1 Stack2
Procedure
Step 1 (Optional) Run the display switchover state command to check whether the stack meets
requirements for a switchover.
Step 2 Run the system-view command to enter the system view.
Step 3 Run the slave switchover enable command to enable master/standby switchover.
By default, master/standby switchover is enabled.
Step 4 Run the slave switchover command to perform a master/standby switchover.
----End
Context
Two methods are available to upgrade stack software: traditional upgrade and smooth
upgrade. Table 9-31 compares the two upgrade methods.
same system
software for the
next startup, with
the same
software package
name, version,
and path.
Procedure
l Traditional upgrade
a. Upload system software to the master switch in the stack. For details on how to
upload files, see File Management in the S1720, S2700, S5700, and S6720
V200R011C10 Configuration Guide - Basic Configuration.
b. Run the startup system-software system-file all command to set the next startup
software for all member switches in the stack.
c. Run the reboot command to upgrade the stack by restarting the entire system.
l Smooth upgrade
a. Upload system software to all the member switches in the stack. Ensure that all
member switches have the same system software, with the same software package
name, version, and path. For details on how to upload files, see File Management in
the S1720, S2700, S5700, and S6720 V200R011C10 Configuration Guide - Basic
Configuration.
b. Run the startup system-software system-file all command to set the next startup
software for all member switches in the stack.
c. Run the system-view command to enter the system view.
d. Run the upgrade backup-area slot slot-id to slot-id command to define the active
and backup areas for a smooth upgrade.
The member switches with stack IDs in the specified range join the backup area,
and the other member switches automatically join the active area.
NOTE
Follow these rules to define the active and backup areas in a stack:
l The active and backup areas cannot have the same member switch, and both areas must
have at least one member switch.
l The backup area cannot contain the master switch of the stack.
l Member switches in each area must be directly connected.
l Member switches in the active and backup areas form the entire stack.
e. (Optional) Run the display upgrade area command to check whether the
configured areas meet the requirements of the current stack topology.
If the stack topology changes after the areas are configured, the smooth upgrade
will fail. To avoid this problem, check the area configuration before starting a
smooth upgrade. If the areas fail the check, define the active and backup areas
correctly according to the current stack topology.
----End
Context
If a stack is no longer used, you can split the stack into standalone switches. A split process
includes the following steps.
1. Back up the stack configuration file.
2. Remove stack cables.
3. Restore the default stack configuration.
4. Power off all the member switches.
5. Remove the stack card. (This step is required only in stack card stacking.)
Procedure
l Splitting a stack established with stack card connections
a. Run the save command to save the configuration.
b. Run the copy source-filename destination-filename all command to back up the
configuration file to all the member switches, so that you can use the configuration
file when you set up a stack the next time.
c. Remove the stack cables between the member switches. After you remove the stack
cables, the stack splits and some member switches restart. The remote login may
fail because some configurations of member switches such as the IP address are the
same. Log in to a switch through the console port to perform configurations.
d. Run the system-view command to enter the system view.
e. Run the stack slot slot-id renumber 0 command to restore the stack ID of a
member switch to the default value 0.
f. Power off all the member switches.
g. Remove the stack card.
l Splitting a stack established through service port connections (using ordinary and
dedicated cables)
a. Run the save command to save the configuration.
b. Run the copy source-filename destination-filename all command to back up the
configuration file to all the member switches, so that you can use the configuration
file when you set up a stack the next time.
c. Remove the stack cables between the member switches. After you remove the stack
cables, the stack splits and some member switches restart. The remote login may
fail because some configurations of member switches such as the IP address are the
same. Log in to a switch through the console port to perform configurations.
d. Run the system-view command to enter the system view.
e. Run the reset stack configuration command to clear the stack configuration.
The cleared stack configuration includes: switch slot ID, stack priority, stack
reserved VLAN, stack MAC address switching delay, stack port configuration, and
stack port rate configuration. Running this command will restart the switch.
f. Power off all the member switches.
----End
Networking Requirements
An enterprise network needs to provide a sufficient number of ports to access devices, but a
simple network structure is required to facilitate configuration and maintenance.
As shown in Figure 9-40, SwitchA, SwitchB, and SwitchC need to set up a stack with a ring
topology and connect to SwitchD through inter-device Eth-Trunks. SwitchA, SwitchB, and
SwitchC are the master, standby, and slave switches respectively, with stack IDs of 0, 1, and 2
and stack priorities of 200, 100, and 100. As they function as a single logical device on the
network, the number of ports increases while keeping network management and maintenance
relatively simple.
The S5720-C-EI series switches are used on the network and can set up a stack using stack
card connections.
Network
SwitchD
Eth-Trunk
Stack link
common link
Eth-Trunk
Configuration Roadmap
The configuration roadmap is as follows:
1. Power off the switches, install one ES5D21VST000 stack card on each switch, and
power on the switches.
2. Configure stack IDs and stack priorities for member switches to facilitate management.
3. Power off SwitchA, SwitchB, and SwitchC. Connect stack ports using QSFP+ cables
according to Figure 9-40 and power on the switches.
4. Configure inter-device Eth-Trunks to increase reliability and uplink bandwidth.
NOTE
When connecting stack cables, connect STACK1 port of one switch to STACK2 port of another switch.
If SwitchA, SwitchB, and SwitchC are different switch models, these models must be able to set up a
stack together. For details, see Licensing Requirements and Limitations for Stacks.
During a stack setup, the software version used by the master switch takes effect, and the standby and
slave switches synchronize their software versions with the software version of the master switch.
Procedure
Step 1 Install the ES5D21VST000 stack card.
Install the ES5D21VST000 stack card on SwitchA, SwitchB, and SwitchC. For details on
how to install a stack card, see 9.9.1 Installing a Stack Card.
Step 2 To help identify devices, set device names as SwitchA, SwitchB, and SwitchC. The following
uses SwitchA as an example.
<HUAWEI> system-view
[HUAWEI] sysname SwitchA
Step 4 Power off SwitchA, SwitchB, and SwitchC, connect stack ports using QSFP+ cables, and
power them on.
NOTE
l Run the save command to save the configurations before you power off the switches.
l The STACK1 port of one switch must be connected to the STACK2 port of another switch.
Otherwise, the stack cannot be set up.
l To ensure that the stack has been successfully established, connect and power on the switches in the
following sequence. (To specify a member switch as the master switch, power on that switch first. In
this example, SwitchA is the master switch.)
1. Power off SwitchA, SwitchB, and SwitchC.
2. Connect SwitchA and SwitchB with a stack cable.
3. Power on and start SwitchA and then power on SwitchB.
4. Check whether SwitchA and SwitchB have set up a stack successfully. For the verification
method, see 9.9.5 Checking Whether a Stack Has Been Established Through Stack Card
Connections.
5. Connect SwitchC to SwitchB and SwitchA using stack cables and then power on SwitchC.
6. Check whether SwitchA, SwitchB, and SwitchC have set up a stack successfully. For the
verification method, see 9.9.5 Checking Whether a Stack Has Been Established Through
Stack Card Connections.
----End
Configuration Files
None
Networking Requirements
An enterprise network needs to provide a sufficient number of ports to access devices, but a
simple network structure is required to facilitate configuration and maintenance.
As shown in Figure 9-41, SwitchA, SwitchB, and SwitchC need to set up a stack with a ring
topology and connect to SwitchD through inter-device Eth-Trunks. SwitchA, SwitchB, and
SwitchC are the master, standby, and slave switches respectively, with stack IDs of 0, 1, and 2
and stack priorities of 200, 100, and 100. As they function as a single logical device on the
network, the number of ports increases while keeping network management and maintenance
relatively simple.
The S5700-LI series switches are used on the network and can set up a stack using service
port connections.
Network
SwitchD
Eth-Trunk
Stack link
common link
Eth-Trunk
Configuration Roadmap
The configuration roadmap is as follows:
1. Configure logical stack ports and add physical member ports to the corresponding logical
stack ports to enable packet forwarding between member switches.
2. Configure stack IDs and stack priorities for member switches to facilitate management.
3. Power off SwitchA, SwitchB, and SwitchC. Connect physical member ports using SFP+
stack cables according to Figure 9-41 and power on the switches.
4. Configure inter-device Eth-Trunks to increase reliability and uplink bandwidth.
NOTE
Physical member ports in stack-port n/1 of one switch can only be connected to the ports in stack-port
m/2 of a neighboring switch.
If SwitchA, SwitchB, and SwitchC are different switch models, these models must be able to set up a
stack together. For details, see Licensing Requirements and Limitations for Stacks.
During a stack setup, the software version used by the master switch takes effect, and the standby and
slave switches synchronize their software versions with the software version of the master switch.
Procedure
Step 1 Configure logical stack ports and add physical member ports to them.
# Configure the service ports GigabitEthernet0/0/27 and GigabitEthernet0/0/28 on SwitchA as
physical member ports and add them to the corresponding logical stack ports.
<HUAWEI> system-view
[HUAWEI] sysname SwitchA
[SwitchA] interface stack-port 0/1
[SwitchA-stack-port0/1] port interface gigabitethernet 0/0/27 enable
Warning: Enabling stack function may cause configuration loss on the interface.
Continue? [Y/N]:y
Info: This operation may take a few seconds. Please wait.
[SwitchA-stack-port0/1] quit
[SwitchA] interface stack-port 0/2
[SwitchA-stack-port0/2] port interface gigabitethernet 0/0/28 enable
Warning: Enabling stack function may cause configuration loss on the interface.
Continue? [Y/N]:y
Info: This operation may take a few seconds. Please wait.
[SwitchA-stack-port0/2] quit
Step 3 Power off SwitchA, SwitchB, and SwitchC, connect them using SFP+ stack cables, and then
power the switches on.
NOTE
l Run the save command to save the configurations before you power off the switches.
l stack-port 0/1 of one switch must be connected to stack-port 0/2 of a neighboring switch. Otherwise,
the stack cannot be set up.
l To ensure that the stack has been successfully established, connect and power on the switches in the
following sequence. (To specify a member switch as the master switch, power on that switch first. In
this example, SwitchA is the master switch.)
1. Power off SwitchA, SwitchB, and SwitchC.
2. Connect SwitchA and SwitchB with a stack cable.
3. Power on and start SwitchA and then power on SwitchB.
4. Check whether SwitchA and SwitchB have set up a stack successfully. For the verification
method, see 9.10.4 Checking Whether a Stack Is Established Through Service Port
Connections.
5. Connect SwitchC to SwitchB and SwitchA using stack cables and then power on SwitchC.
6. Check whether SwitchA, SwitchB, and SwitchC have set up a stack successfully. For the
verification method, see 9.10.4 Checking Whether a Stack Is Established Through Service
Port Connections.
----End
Configuration Files
None
Networking Requirements
An enterprise network needs to provide sufficient ports for access devices, and the network
structure should be simple to facilitate configuration and management.
As shown in Figure 9-42, Switches A to C set up a stack in a ring topology and connect to
SwitchD through an inter-chassis Eth-Trunk. To reduce the configuration, Switches A to C set
up a stack using dedicated stack cables. In the stack, SwitchA needs to function as the master
switch.
Network
SwitchD
Eth-Trunk
Stack link
common link
Eth-Trunk
Configuration Roadmap
The configuration roadmap is as follows:
1. Power off SwitchA, SwitchB, and SwitchC to ensure security.
2. Connect the switches using dedicated stack cables based on dedicated stack cable
connection rules.
3. Power on SwitchA first to ensure that it becomes the master switch.
4. Save the stack configuration automatically generated for dedicated cable stacking to the
flash memory. This ensures that the stack configuration still takes effect when these
cables are removed or other cables are connected.
5. Configure an inter-chassis Eth-Trunk to increase reliability and uplink bandwidth.
Procedure
Step 1 Power off SwitchA, SwitchB, and SwitchC.
Step 2 Connect SwitchA, SwitchB, and SwitchC using dedicated stack cables as shown in Figure
9-43. Ensure that logical stack port 1 of the local switch is connected to logical stack port 2 of
the adjacent switch. Otherwise, these switches cannot set up a stack.
SwitchA
Logical stack port 1 GE0/0/27
Logical stack
GE0/0/26 port 2
SwitchB
Logical stack port 1 GE0/0/27
Logical stack
GE0/0/26 port 2
SwitchC
Step 5 Save the stack configuration automatically generated for dedicated cable stacking to the flash
memory.
# After checking that a stack has been set up, save the stack configuration automatically
generated for dedicated cable stacking to the flash memory.
----End
Configuration files
None
Networking Requirements
As shown in Figure 9-44, SwitchA and SwitchB set up a stack. The stack IDs of SwitchA and
SwitchB are 0 and 1 respectively.
MAD can be used to detect dual master switches with the same configuration on the network
to reduce the impact of a stack split on the network.
SwitchC
Stack
SwitchA SwitchB
GE0/0/5 GE1/0/5
MAD Link
Stack Link
Common Link
Configuration Roadmap
The configuration roadmap is as follows:
Procedure
Step 1 Configure MAD in direct mode on a specified port.
If no service is configured in the stack, perform the following steps to check whether the
MAD function has been configured successfully:
2. Simulate a stack split situation by removing some stack cables. If the stack is set up
through service port connections, you can also run the shutdown interface command to
shut down some physical member ports in the logical stack port view.
3. After the stack splits, you can see the following alarm on the terminal: MAD/4/
MULTIACTIVEDETECTED(t):OID 1.3.6.1.4.1.2011.5.25.246.1.1 Multi-active
scenario is detected.
4. Check whether the number of member switches decreases.
Run the display device command, and you can see that the number of member switches
decreases from 2 to 1. (In this example, the member switches are S5700-28P-LI-AC.)
<HUAWEI> display device
S5700-28P-LI-AC's Device status:
Slot Sub Type Online Power Register Status Role
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
-
0 - S5700-28P-LI Present PowerOn Registered Normal
Master
5. The preceding operations verify that the MAD function has been configured
successfully.
6. Connect the stack cables again or run the undo shutdown command to enable the
physical member ports, so that the stack can be set up again.
7. After the stack is set up again, you can see the following alarm on the terminal: MAD/4/
MULTIACTIVEMERGE(t):OID 1.3.6.1.4.1.2011.5.25.246.1.2 Multi-active scenario
is merged.
8. Check whether the number of member switches restores to the original value.
Run the display device command, and you can see that the number of member switches
changes to 2. (In this example, the member switches are S5700-28P-LI-AC.)
<HUAWEI> display device
S5700-28P-LI-AC's Device status:
Slot Sub Type Online Power Register Status Role
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
-
0 - S5700-28P-LI Present PowerOn Registered Normal
Master
1 - S5700-28P-LI Present PowerOn Registered Normal
Standby
----End
Configuration Files
l Stack configuration file
#
interface GigabitEthernet0/0/5
mad detect mode direct
#
interface GigabitEthernet1/0/5
mad detect mode direct
#
return
Configuration Requirements
As shown in Figure 9-45, SwitchA, SwitchB, and SwitchC set up a stack and they connect to
SwitchD through Eth-Trunk1.
MAD can be used to detect dual master switches with the same configuration on the network
to reduce the impact of a stack split on the network.
Network
SwitchD
GE0/0/1 GE0/0/3
GE0/0/2
Eth-Trunk1
MAD Link
Stack Link
Common Link
Configuration Roadmap
The configuration roadmap is as follows:
1. On the stack, configure MAD in relay mode on the inter-device Eth-Trunk.
2. On the relay agent, configure the MAD relay function so that MAD packets can be
forwarded through the Eth-Trunk.
Procedure
Step 1 On the stack, configure MAD in relay mode on the inter-device Eth-Trunk.
<HUAWEI> system-view
[HUAWEI] interface eth-trunk 1
[HUAWEI-Eth-Trunk1] mad detect mode relay
[HUAWEI-Eth-Trunk1] quit
[HUAWEI] interface gigabitethernet 0/0/5
[HUAWEI-GigabitEthernet0/0/5] eth-trunk 1
[HUAWEI-GigabitEthernet0/0/5] quit
[HUAWEI] interface gigabitethernet 1/0/5
[HUAWEI-GigabitEthernet1/0/5] eth-trunk 1
[HUAWEI-GigabitEthernet1/0/5] quit
[HUAWEI] interface gigabitethernet 2/0/5
[HUAWEI-GigabitEthernet2/0/5] eth-trunk 1
[HUAWEI-GigabitEthernet2/0/5] quit
[HUAWEI] quit
<HUAWEI> system-view
[HUAWEI] sysname SwitchD
[SwitchD] interface eth-trunk 1
[SwitchD-Eth-Trunk1] mad relay
[SwitchD-Eth-Trunk1] quit
[SwitchD] interface gigabitethernet 0/0/1
[SwitchD-GigabitEthernet0/0/1] eth-trunk 1
[SwitchD-GigabitEthernet0/0/1] quit
[SwitchD] interface gigabitethernet 0/0/2
[SwitchD-GigabitEthernet0/0/2] eth-trunk 1
[SwitchD-GigabitEthernet0/0/2] quit
[SwitchD] interface gigabitethernet 0/0/3
[SwitchD-GigabitEthernet0/0/3] eth-trunk 1
[SwitchD-GigabitEthernet0/0/3] quit
[SwitchD] quit
2. Simulate a stack split situation by removing the stack cables that connect SwitchB to
SwitchC and SwitchC to SwitchA. If the stack is set up through service port connections,
you can also run the shutdown interface command to shut down the physical member
ports in the logical stack port view that connect SwitchB to SwitchC and SwitchC to
SwitchA.
3. After the stack splits, you can see the following alarm on the terminal: MAD/4/
MULTIACTIVEDETECTED(t):OID 1.3.6.1.4.1.2011.5.25.246.1.1 Multi-active
scenario is detected.
5. The preceding operations verify that the MAD function has been configured
successfully.
6. Connect the stack cables again or run the undo shutdown command to enable the
physical member ports, so that the stack can be set up again.
7. After the stack is set up again, you can see the following alarm on the terminal: MAD/4/
MULTIACTIVEMERGE(t):OID 1.3.6.1.4.1.2011.5.25.246.1.2 Multi-active scenario
is merged.
8. Check whether the number of member switches restores to the original value.
Run the display device command, and you can see that the number of member switches
changes to 3. (In this example, the member switches are S5700-28P-LI-AC.)
<HUAWEI> display device
S5700-28P-LI-AC's Device status:
Slot Sub Type Online Power Register Status Role
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
-
0 - S5700-28P-LI Present PowerOn Registered Normal
Master
1 - S5700-28P-LI Present PowerOn Registered Normal
Standby
2 - S5700-28P-LI Present PowerOn Registered Normal Slave
----End
Configuration Files
l Stack configuration file
#
interface Eth-Trunk1
mad detect mode relay
#
interface GigabitEthernet0/0/5
eth-trunk 1
#
interface GigabitEthernet1/0/5
eth-trunk 1
#
interface GigabitEthernet2/0/5
eth-trunk 1
#
return
#
interface GigabitEthernet0/0/1
eth-trunk 1
#
interface GigabitEthernet0/0/2
eth-trunk 1
#
interface GigabitEthernet0/0/3
eth-trunk 1
#
return
Network
SwitchE
GE0/0/1 GE0/0/2
Eth-Trunk 10
SwitchA GE
SwitchB
1/0 /5
GE1/0/3 /5 2/0
GE GE2/0/3
Eth-Trunk 20 Eth-Trunk 30
GE0/0/2 GE0/0/2
GE0/0/1 GE0/0/1
SwitchC SwitchD
Stack link
common link
Eth-Trunk
Configuration Roadmap
The configuration roadmap is as follows:
1. Configure a stack Eth-Trunk between the stack and its upstream device and add physical
member ports to the Eth-Trunk to expand the uplink bandwidth.
2. Configure stack Eth-Trunks between the stack and its downstream devices and add
physical member ports to the Eth-Trunks, so that the member switches work in
redundancy mode to improve network reliability.
3. Enable Eth-Trunks to forward traffic from local ports first to improve forwarding
efficiency and reduce the load on the stack cable between member switches. When an
Eth-Trunk member port of a local device is working normally or when the traffic is not
heavy, traffic is forwarded preferentially through the local member port.
Procedure
Step 1 Configure an Eth-Trunk between the stack and its upstream device.
# Configure an Eth-Trunk on the stack and add uplink ports to the Eth-Trunk.
<HUAWEI> system-view
[HUAWEI] sysname Stack
[Stack] interface eth-trunk 10
[Stack-Eth-Trunk10] quit
[Stack] interface gigabitethernet 1/0/4
[Stack-GigabitEthernet1/0/4] eth-trunk 10
[Stack-GigabitEthernet1/0/4] quit
[Stack] interface gigabitethernet 2/0/4
[Stack-GigabitEthernet2/0/4] eth-trunk 10
[Stack-GigabitEthernet2/0/4] quit
Step 2 Configure Eth-Trunks between the stack and its downstream devices.
# Configure an Eth-Trunk on the stack and add the downlink ports connected to SwitchC to
the Eth-Trunk.
[Stack] interface eth-trunk 20
[Stack-Eth-Trunk20] quit
[Stack] interface gigabitethernet 1/0/3
[Stack-GigabitEthernet1/0/3] eth-trunk 20
[Stack-GigabitEthernet1/0/3] quit
[Stack] interface gigabitethernet 2/0/5
[Stack-GigabitEthernet2/0/5] eth-trunk 20
[Stack-GigabitEthernet2/0/5] quit
# Configure an Eth-Trunk on the stack and add the downlink ports connected to SwitchD to
the Eth-Trunk.
[Stack] interface eth-trunk 30
[Stack-Eth-Trunk30] quit
[Stack] interface gigabitethernet 1/0/5
[Stack-GigabitEthernet1/0/5] eth-trunk 30
[Stack-GigabitEthernet1/0/5] quit
[Stack] interface gigabitethernet 2/0/3
[Stack-GigabitEthernet2/0/3] eth-trunk 30
[Stack-GigabitEthernet2/0/3] quit
Step 3 Enable local preferential forwarding on the Eth-Trunks. By default, local preferential
forwarding is enabled on an Eth-Trunk.
[Stack] interface eth-trunk 10
[Stack-Eth-Trunk10] local-preference enable
[Stack-Eth-Trunk10] quit
[Stack] interface eth-trunk 20
[Stack-Eth-Trunk20] local-preference enable
[Stack-Eth-Trunk20] quit
[Stack] interface eth-trunk 30
[Stack-Eth-Trunk30] local-preference enable
[Stack-Eth-Trunk30] quit
[Stack] quit
----End
Configuration Files
l Stack configuration file
#
sysname Stack
#
interface Eth-Trunk10
#
interface Eth-Trunk20
#
interface Eth-Trunk30
#
interface GigabitEthernet1/0/3
eth-trunk 20
#
interface GigabitEthernet1/0/4
eth-trunk 10
#
interface GigabitEthernet1/0/5
eth-trunk 30
#
interface GigabitEthernet2/0/3
eth-trunk 30
#
interface GigabitEthernet2/0/4
eth-trunk 10
#
interface GigabitEthernet2/0/5
eth-trunk 20
#
return
Networking Requirements
As shown in Figure 9-47, SwitchA, SwitchB, and SwitchC set up a stack to support user
access and connect to an upstream device SwitchD through an Eth-Trunk. SwitchA, SwitchB,
and SwitchC are the master, standby, and slave switches respectively, with stack IDs of 0, 1,
and 2 and stack priorities of 200, 100, and 100. As an increasing number of users need to
connect to the network, the current stack cannot provide a sufficient number of ports.
SwitchE is added to the stack as shown in Figure 9-48 to increase the number of ports
available for user access.
The S5700-LI series switches are used on the network and can set up a stack using service
port connections.
Network
SwitchD
Eth-Trunk
Stack link
common link
Eth-Trunk
Network
SwitchD
Eth-Trunk
Stack link
common link
Eth-Trunk
Configuration Roadmap
The configuration roadmap is as follows:
1. Examine the physical connections of the stack to determine a location for the new
member switch.
2. Perform software configurations for service port stacking on SwitchE: configure logical
stack ports and add physical member ports to the corresponding logical stack ports to
enable packet forwarding between member switches, and configure stack IDs to facilitate
user management.
3. Power off SwitchE, connect physical member ports using SFP+ stack cables according to
Figure 9-48, and then power on SwitchE.
NOTE
Physical member ports of logical stack port stack-port n/1 on one switch can only be connected to the
ports of stack-port m/2 on a neighboring switch.
4. Add the uplink port of SwitchE to the inter-device Eth-Trunk.
Procedure
Step 1 Determine a location for the new member switch.
# The stack in this example has a ring topology. Break the ring topology into a chain topology
and add the new member switch to either end of the chain. Disconnect the link between
SwitchA and SwitchC to break the ring.
NOTE
If the member switches are running a software version earlier than V200R001, break the ring on the link
between the master switch and standby switch. Otherwise, reelection will occur which will cause unnecessary
role changes.
NOTE
The physical member ports GigabitEthernet0/0/27 and GigabitEthernet0/0/28 on SwitchA, SwitchB, and
SwitchC are added to the logical stack ports stack-port n/1 and stack-port n/2 respectively.
<HUAWEI> system-view
[HUAWEI] sysname SwitchE
[SwitchE] interface stack-port 0/1
[SwitchE-stack-port0/1] port interface gigabitethernet 0/0/27 enable
Warning: Enabling stack function may cause configuration loss on the interface.
Continue? [Y/N]:y
Info: This operation may take a few seconds. Please wait.
[SwitchE-stack-port0/1] quit
[SwitchE] interface stack-port 0/2
[SwitchE-stack-port0/2] port interface gigabitethernet 0/0/28 enable
Warning: Enabling stack function may cause configuration loss on the interface.
Continue? [Y/N]:y
Info: This operation may take a few seconds. Please wait.
[SwitchE-stack-port0/2] quit
Step 3 Power off SwitchE, connect physical member ports on SwitchE to corresponding physical
member ports on SwitchA and SwitchC according to Figure 9-48, and then power on
SwitchE.
NOTE
It is recommended that you run the save command to save the configurations before you power off
SwitchE.
----End
Configuration Files
None
Networking Requirements
In Figure 9-49, SwitchA, SwitchB, and SwitchC set up a stack using dedicated stack cables to
provide access service to users. The stack connects to SwitchD through an inter-chassis Eth-
Trunk. SwitchA, SwitchB, and SwitchC are the master, standby, and slave switches
respectively, with stack IDs of 0, 1, and 2 respectively. As the number of users increases, the
current stack cannot provide sufficient ports for user access.
SwitchE needs to be added to the stack to increase the number of ports. Figure 9-50 shows
the topology of the new stack.
Figure 9-49 Before the new member switch is added to the stack
Network
SwitchD
Eth-Trunk
Stack link
common link
Eth-Trunk
Figure 9-50 After the new member switch is added to the stack
Network
SwitchD
Eth-Trunk
Stack link
common link
Eth-Trunk
Configuration Roadmap
The configuration roadmap is as follows:
1. Analyze physical connections between original stack member switches to determine
where to add the new member switch.
2. Power off SwitchE. Connect stack member ports using dedicated stack cables according
to Figure 9-51 and then power on the switches.
NOTE
Procedure
Step 1 Determine the location for the new member switch.
# The stack in this example has a ring topology. To add the new member switch to the stack,
tear down the link between SwitchA and SwitchC to change the stack topology to a chain
topology, and then connect SwitchE to SwitchA and SwitchC.
Step 2 Power off SwitchE, use dedicated stack cables to connect GE0/0/27 and GE0/0/26 on
SwitchE to corresponding ports on SwitchA and SwitchC according to Figure 9-51, and then
power on SwitchE.
SwitchA
Logical stack port 1 GE0/0/27
Logical stack
GE0/0/26 port 2
SwitchB
Logical stack port 1 GE0/0/27
Logical stack
GE0/0/26 port 2
SwitchC
SwitchE
----End
Configuration Files
None
Networking Requirements
As shown in Figure 9-52, switches A to C set up a stack to provide access service to users.
The stack connects to SwitchD through an inter-chassis Eth-Trunk. SwitchA, SwitchB, and
SwitchC are the master, standby, and slave switches respectively, with stack IDs of 0, 1, 2 and
stack priorities of 200, 100, 100, respectively. As the number of users increases, the current
stack cannot provide sufficient ports for user access.
SwitchE needs to be added to the stack to increase the number of ports. Figure 9-53 shows
the topology of the new stack.
In this example, the stacked switches are S5720-EI, and they are connected using stack cards.
Figure 9-52 Before the new member switch is added to the stack
Network
SwitchD
Eth-Trunk
Stack link
Common link
Eth-Trunk
Figure 9-53 After the new member switch is added to the stack
Network
SwitchD
Eth-Trunk
Stack link
Common link
Eth-Trunk
Configuration Roadmap
The configuration roadmap is as follows:
1. Analyze physical connections between original stack member switches to determine
where to add the new member switch.
2. Set a stack ID for SwitchE.
3. Power off SwitchE. Use QSFP+ high-speed copper cables to connect physical stack
member ports according to Figure 9-53, and then power on SwitchE.
NOTE
Procedure
Step 1 Determine the location for the new member switch.
# The stack in this example has a ring topology. To add the new member switch to the stack,
tear down the link between SwitchA and SwitchC to change the stack topology to a chain
topology, and then connect SwitchE to SwitchA and SwitchC.
NOTE
If the member switches are running a software version earlier than V200R001, do not tear down the link
between the master and standby switches, as this will result in reelection in the stack.
Step 3 Power off SwitchE, use QSFP+ high-speed copper cables to connect physical stack member
ports on SwitchE to corresponding physical member ports on SwitchA and SwitchC
according to Figure 9-53, and then power on SwitchE.
NOTE
Run the save command to save the configuration before you power off SwitchE.
----End
Configuration Files
None
Networking Requirements
A stack is running stably and transmitting upstream and downstream services normally. The
software of stacked switches needs to be upgraded without interrupting upstream and
downstream services.
As shown in Figure 9-54, switches A to C set up a stack in ring topology and connect to
SwitchD through an inter-chassis Eth-Trunk. SwitchA, SwitchB, and SwitchC are the master,
standby, and slave switches of the stack, with stack IDs of 0, 1, and 2, respectively.
In this example, the stacked switches are S5700-LI, and they are connected through service
ports.
Network
SwitchD
Eth-Trunk
Stack link
common link
Eth-Trunk
Configuration Roadmap
The configuration roadmap is as follows:
1. Upload the target system software package to all stack member switches.
2. Define the active and backup areas for a smooth upgrade.
3. Start the smooth upgrade.
Procedure
Step 1 Upload the target system software package to all stack member switches.
Ensure that all member switches have the same system software, with the same software
package name, version, and path. For details on how to upload files to a switch, see File
Management in the S1720, S2700, S5700, and S6720 V200R011C10 Configuration Guide -
Basic Configuration.
Step 2 Define the active and backup areas for a smooth upgrade.
# Assign SwitchB and SwitchC to the backup area, and SwitchA to the active area.
<HUAWEI> startup system-software test.cc all
<HUAWEI> system-view
[HUAWEI] upgrade backup-area slot 1 to 2
The member switches with stack IDs in the specified range join the backup area, and the other
member switch automatically joins the active area.
NOTE
Follow these rules to define the active and backup areas in a stack:
l The active and backup areas cannot have the same member switch, and both areas must have at
least one member switch.
l The backup area cannot contain the master switch.
l Member switches in each area must be directly connected.
l Member switches in the active and backup areas form the entire stack.
----End
Configuration Files
None
l A stack cannot be set up among different switch series. For example, a stack cannot be
set up among the S5700 and S6700 switches.
l A stack cannot be set up among different switch models of the same switch series. For
example, a stack cannot be set up among the S5720EI and S5720HI switches.
l A stack can be set up among only certain different switch models of the same switch
series. For example, a stack can be set up among the S5720SI and S5720S-SI switches
but not among the S5720LI and S5720S-LI switches. For details about whether a stack
can be set up among different switch models, see stack support of different switch
models in Stacking Support and Version Requirements. As described in Remarks of
S6720LI and S6720S-LI Service Port Stacking Support, a stack can be set up among
all S6720LI models, among all S6720S-LI models, but not among the S6720LI and
S6720S-LI models.
For details about how to apply for a license, see S Series Switch License Use
Guide.
NOTE
In V200R009 and later versions, if all member switches have installed license files with the same
license-controlled items before they set up a stack, you do not need to apply for a new license for the
stack. If only some of member switches have installed a license file, apply for a new license for the
stack.
In versions prior to V200R009, apply for a license that contains ESNs of all the member switches.
2. Install the license file to the stack master switch.
3. Run the license active command to activate the license file. The standby and slave
switches will synchronize the license from the master switch.
4. After the license file is installed successfully, run the display license command to check
license file information.
After the stack splits, the license is still effective on all the member switches.
l You must have a Huawei account to log in to the license self-service system. For details
on how to apply for a Huawei account, click Help, and then click Applying for an
Account under the Basic Concepts node.
l The following is the license processing mechanism in a stack system. Knowing this will
help you handle license status exceptions.
– If member switches are running a version prior to V200R009, you must apply for a
license that contains ESNs of all member switches. If the license file contains only
the ESN of the master switch, the license is in normal state and the license-
controlled features can take effect in the stack. If a switchover occurs in the stack,
however, the license will change to the trial state (60-day trial period). If the license
file contains only the ESN of the standby switch or a slave switch, the license
remains in trial state after it is loaded.
– If member switches are running V200R009 or a later version, you can apply for a
license that contains ESNs of all member switches, or apply a license for each ESN
and install the licenses to the switches before setting up a stack. When using the
second method, ensure that all license files have the same license-controlled items,
because:
n If the license files of the standby and slave switches have the same items as the
license file of the master switch, the standby and slave switches use their own
license files and do not synchronize the license file from the master switch.
Their license files are still effective after a switchover in the stack.
n If the license files of the standby and slave switches have different items than
the license file of the master switch, the standby and slave switches
synchronize the license file from the master switch. If the license file of the
master switch contains the ESNs of the standby and slave switches, the license
is in normal state on the standby and slave switches and will still effective after
a switchover. If the license file on the master switch does not contain the ESNs
of the standby and slave switches, the license will change to trial state (60-day
trial period) on the standby and slave switches, because ESN in the
synchronized license file is different from the ESN of the local switch. After a
switchover, the license of the stack will also change to trial state.
Run the display switchover state command to view the stack master/standby status.
<HUAWEI> display switchover state
Slot 1 HA FSM State(master): realtime or routine backup
Slot 2 HA FSM State(slave): receiving realtime or routine data.
When the master status displays realtime or routine backup and the slave status displays
receiving realtime or routine data, the standby switch does not restart and automatically
becomes the new master switch if the master switch is powered off and restarts. When the
master status and slave status display other states, the standby switch also restarts if the
master switch is powered off and restarts.