Clustered Arcserve Backup servers provide service through virtual name and support backup job failover capability. When the active Arcserve Backup server in a cluster fails, these backup jobs are dispersed from the failed server to other Arcserve Backup servers in the cluster. After Arcserve Backup services are resumed in another cluster node, any failed jobs from the previous server are rerun in a new cluster node.
Arcserve Backup HA server supports two types of failovers; planned failovers and unplanned failovers.
Planned failovers occur when it is necessary to perform maintenance on the active node within a cluster and you want Arcserve Backup to migrate the cluster resources from that active node to a passive node within the cluster. Examples for planned failovers are system maintenance, disaster recovery tests, and training.
When a planned failover occurs, Arcserve Backup recovers in another node with all scheduled jobs retained.
Unplanned failover can occur because of hardware or software failures. When unplanned failover occurs, Arcserve Backup recovers in another node, picks up the failed job from the Arcserve Backup job queue, and resumes the job from the point where it failed. If a failover occurs, the job resume is based on a checkpoint mechanism as follows:
For example, if a backup job involves two volumes: C and D, and a failover occurs when the backup for volume C was finished and the backup for volume D was ongoing. After the failover, the backup job will restart and skip the backup for volume C and continue to backup volume D.
For example, if a backup job involves Host1 and Host2, and a failover occurs when the backup for Host1 is finished, but the backup for Host2 is not. After the failover, the backup job will restart and skip the backup for Host1 and continue to backup Host2 (in this case, the backup for Host2 does not skip any volumes that might be backed up before failover).
Jobs running on other backup servers instead of the HA server of the domain will rarely be impacted by failover. For example, when the Primary server is HA server, and it fails over, the jobs running on member servers are not impacted except in one situation. If you are using a HA primary server, the jobs that run on the member servers may fail when an unplanned failover occurs on the HA primary server. (The failures only occur when the jobs on the member servers are finishing when the failover occurs)
Note: If you are using Arcserve Backup Agents to backup the active node of the cluster or the virtual node, and an unplanned failover occurs (the active node is down), the job would become incomplete. To ensure that these nodes could be backed up after failover, you should configure the jobs to create makeup jobs.
Copyright © 2016 |
|