Information About GIR
You can use Graceful Insertion and Removal (GIR) to put a switch in maintenance mode in order to perform debugging or an upgrade. When switch maintenance is complete, you can return the switch to normal mode.
When you place the switch in maintenance mode, all protocols are isolated from the network. When normal mode is restored, all the protocols are brought back up.
-
Border Gateway Protocol (BGP)
-
BGPv6
-
Enhanced Interior Gateway Routing Protocol (EIGRP)
-
EIGRPv6
-
Intermediate System-to-Intermediate System (ISIS)
-
ISISv6
-
Open Shortest Path First (OSPF)
-
OSPFv3
-
RIP
-
Virtual port channel (vPC) switches
Note
GIR is not supported on vPC+ switches.
-
Interfaces
-
FabricPath
-
Border Gateway Protocol (BGP)
-
BGPv6
-
Enhanced Interior Gateway Routing Protocol (EIGRP)
-
EIGRPv6
-
Intermediate System-to-Intermediate System (ISIS)
-
ISISv6
-
Open Shortest Path First (OSPF)
-
OSPFv3
-
FabricPath (Only applicable for Spine switches)
Note |
|
Maintenance Profile
Maintenance profile contains a set of commands that will be applied sequentially during graceful removal or graceful insertion.
By default, the system isolates all enabled protocols during graceful removal and restores them during graceful insertion. The protocols are isolated and restored in a predefined order.
The switch supports the following profiles:
-
Maintenance-mode profile—Contains all the commands that will be executed during graceful removal, when the switch enters maintenance mode.
-
Normal-mode profile—Contains all the commands that will be executed during graceful insertion, when the switch returns to normal mode.
System-generated Profile
You can allow the system to generate a maintenance-mode or normal-mode profile with specific configuration commands. The system generates a maintenance-mode profile when you use the system mode maintenance command or a normal-mode profile when you use the no system mode maintenance command.
Custom Profile
You can create a custom maintenance-mode or normal-mode profile with configuration commands that can be applied during graceful removal or graceful insertion to isolate, shut down, or restore the protocols individually (or perform additional configurations). You can use a custom profile when the system-generated profile does not provide the required configuration or if you need to enhance the existing system-generated or custom profile to include additional functionality specific to your deployment. Use the configure maintenance profile maintenance-mode command to configure a custom maintenance-mode profile with the required commands or the configure maintenance profile normal-mode command to configure a custom normal-mode profile with the required commands.
Note |
We recommend using Scenario 1 or 2. |
Unplanned Maintenance
You can put the switch in unplanned maintenance mode when the switch reloads due to a critical failure. For switches with a single supervisor, configure a reset reason CLI using the system mode maintenance on-reload reset-reason command to enable the switch to go into maintenance mode after a switch reloads due to a critical failure. For switches with dual supervisors, SUP switchover occurs when there is a critical failure of the switch and the switch will not go into maintenance mode. The maintenance-mode profile existing in the startup configuration is applied when the switch goes in to unplanned maintenance mode. If no maintenance mode profile exists in the startup configuration, a system-generated maintenance-mode profile is created and applied when the switch goes in to unplanned maintenance mode.
Maintenance Mode Timer
Use the system mode maintenance timeout command before entering maintenance mode to keep the switch in maintenance mode for a specified number of minutes. You can also use this command while the switch is in maintenance mode to change the number of minutes for which the switch will be in maintenance mode. The timer will then restart from that instant with the new timer value. Once the configured time elapses, the switch returns to normal mode automatically without using the no system mode maintenance mode command. Use the no system mode maintenance timeout command to disable the timer.
Snapshot
Use the snapshot command to capture the running states of selected features and to store the running states on the persistent storage media.
You can use snapshots to compare the state of a switch before it went into maintenance mode and after it came back to normal mode. The snapshot process consists of three parts:
-
Creating a snapshot of the states of a few preselected features on the switch and storing them on the persistent storage media.
-
Listing the snapshots taken at various time intervals and managing them.
-
Comparing snapshots and showing the summary and details of each feature.
There are two types of snapshots:
-
System-generated snapshot—This is generated by the system when you use the [no] system mode maintenance command. The system creates the before_maintenance snapshot just before the system goes into maintenance mode. The system creates the after_maintenance snapshot just before the system goes into normal mode. The system overwrites any old snapshots when you use the [no] system mode maintenance command. Use the snapshot delete { all | snapshot-name } command to delete the system-generated snapshots.
In certain scenarios, the system-generated after_maintenance snapshot may be taken when hardware programming is ongoing. In such cases, we recommend taking a user-generated snapshot after the system has completed hardware programming and is in a stable state. You can then compare the new after_maintenance snapshot with the before_maintenance snapshot.
Starting with Cisco NX-OS release 8.0(1), the [no] system mode maintenance command has been enhanced to execute a normal mode profile and activate a timer ensuring that sufficient time is provided for the switch to complete any hardware programming that may be going on before the after_maintenance snapshot is taken. Once the timer expires, the after_maintenance snapshot is taken in the background and a new warning syslog message, MODE_SNAPSHOT_DONE, is sent after the snapshot is complete. The default delay timer value is 120 seconds. The output of the [no] system mode maintenance command displays the delay timer value, in seconds, after which the after_maintenance snapshot is generated:
The after_maintenance snapshot will be generated in <delay> seconds After that time, please use 'show snapshots compare before_maintenance after_maintenance' to check the health of the system
Use the system mode maintenance snapshot-delay [delay-in-seconds] command to change the delay timer value. The delay timer range is from 0 to 65535.
Use the show maintenance snapshot-delay command to display the snapshot delay timer value.
-
User-generated snapshot—Use the snapshot create name description command to create a user-generated snapshot. Use thesnapshot delete { all | snapshot-name } command to delete user-generated snapshots.
The system-generated and user-generated snapshots are stored in the snapshot repository.
The following table lists the snapshot sections with the corresponding show commands:
Name of the Section |
Corresponding ‘show’ command |
---|---|
bgp-sessions |
show bgp sessions vrf all |
eigrp |
show ip eigrp topology summary |
eigrpv6 |
show ipv6 eigrp topology summary |
interface |
show interface |
ospf |
show ip ospf vrf all |
ospfv3 |
show ipv6 ospfv3 vrf all |
isis |
show isis database detail vrf all |
rip |
show ip rip vrf all |
route-summary |
show ip route summary vrf all |
routev6-summary |
show ipv6 route summary vrf all |
vpc |
show vpc |
Suppress FIB Pending
The Suppress Forwarding Information Base (FIB) Pending feature uses the Border Gateway Protocol-Routing Information Base (BGP-RIB) and the Enhanced Interior Gateway Routing Protocol-Routing Information Base (EIGRP-RIB) feedback mechanism to avoid premature route advertisements and subsequent packet loss in a network. This mechanism is enabled by default and ensures that routes are installed locally before they are advertised to a neighbor.
BGP and EIGRP wait for feedback from RIB indicating that the routes that EIGRP or BGP installed in the RIB are installed in the FIB before EIGRP or BGP sends out updates to the neighbors. EIGRP or BGP will send out updates of only those routes that have versions up to the version that FIB has installed. This selective update ensures that EIGRP or BGP does not send out premature updates resulting in attracting traffic even before the data plane is programmed after a switch reload, line card reload, or when the switch moves to normal mode from maintenance mode.
GIR SNMP Traps
Starting with Cisco NX-OS Release 8.0(1), support for Simple Network Management Protocol (SNMP) traps has been added to the Graceful Insertion and Removal (GIR) mechanism. You can enable the switch to send an SNMP trap notification when the switch moves from normal mode to maintenance mode and vice-versa. Use the snmp-server enable traps mmode cseMaintModeChangeNotify command to enable the switch to send an SNMP trap notification when the switch moves to maintenance mode. Use the snmp-server enable traps mmode cseNormalModeChangeNotify command to enable the switch to send an SNMP trap notification when the switch moves to normal mode. By default, both SNMP traps are disabled.