Specifications

System Redundancy
ExtremeWare XOS 11.0 Concepts Guide 49
Relinquishing Master Status
You can cause the master to failover to the backup, thereby relinquishing its master status. To cause the
failover, complete the following steps:
1 Use the show switch {detail} command to confirm that the nodes are synchronized and have
identical software and switch configurations before failover. The output displays the status of the
MSMs, with the master MSM showing
MASTER and the backup MSM showing BACKUP (InSync).
A node may not be synchronized because checkpointing did not occur, incompatible software is
running on the master and backup, or the backup is down.
• If the nodes are not synchronized, and both MSMs are running ExtremeWare XOS 11.0 or later,
proceed to step 2.
• If the nodes are not synchronized, and one MSM is running ExtremeWare XOS 10.1 or earlier,
proceed to step 3.
• If the nodes are synchronized, proceed to step 3.
2 Use the synchronize command to ensure that the backup has the same software in flash as the
master.
NOTE
Both the backup and the master MSMs must be running ExtremeWare XOS 11.0 or later to use the
synchronize command.
The
synchronize command:
• Reboots the backup MSM to prepare it for synchronizing with the master MSM
• Copies both the primary and secondary software images
• Copies both the primary and secondary configurations
• Reboots the backup MSM after replication is complete
3 Initiate failover from the master MSM to the backup MSM.
• If both nodes are running ExtremeWare XOS 11.0 or later, use the run msm-failover command.
• If one node is running ExtremeWare XOS 10.1 or earlier, use the
run msm-failover {force}
command. By specifying
force, failover occurs regardless of the version of software running on
the MSMs.
Replicating Data Between Nodes
ExtremeWare XOS replicates configuration and run-time information between the master MSM and the
backup MSM so that the system can recover if the master fails. This method of replicating data is
known as checkpointing. Checkpointing is the process of automatically copying the active state from the
master to the backup, which allows for state recovery if the master fails.
Replicating data consists of the following three steps:
1 Configuration synchronization—Relays current and saved configuration information from the master
to the backup
2 Bulk checkpoint—Ensures that each individual application running on the system is synchronized
with the backup
3 Dynamic checkpoint—Checkpoints any new state changes from the master to the backup
To monitor the checkpointing status, use the show checkpoint-data {<process>} command.