Home > Storage > PowerMax and VMAX > Storage Admin > Implementing Dell SRDF SRA with VMware SRM > Known issues with AutoTargetDeviceReuse=YES
The following sections cover special considerations or restrictions when setting the previously covered parameter AutoTargetDeviceReuse=YES.
As customer environments grow, and datastores or RDMs are added, the composition of an individual protection group as well as the underlying device/composite group will need to be updated. In the main, this is a straightforward process of adding the devices to said groups, and then running a rediscovery of pairs in SRM. Protection groups and recovery plans can then be adjusted if required.
While these steps are sufficient for planned migration or failover, special care should be taken if test failover was in use prior to adding the devices. In particular, for those customers utilizing AutoTargetDeviceReuse=YES, adding new device pairs can cause undesireable expansion of the EmcSrdfSraTestFailoverConfig.xml file by the SRA when the next failover test is run. The reason for this is that if snapshots exist from a previous test, the SRA will not modify the existing <CopyInfo> block, rather it will add a new block for the new device pair(s). To avoid this condition, it is important before adding new device pairs to run a test failover with AutoTargetDeviceReuse=NO. This will effectively clean the EmcSrdfSraTestFailoverConfig.xml file so that after the new device pair(s) are added, there will be a single <CopyInfo> block.
Note: Although this documented process is a best practice, it will not prevent test failover from working after the additional pairs are added to the environment with AutoTargetDeviceReuse=YES.
VMware supports the execution of SRM recovery plans simultaneously. This can be particularly advantageous when a true failover event occurs as it can reduce the recovery time objective. If customers plan on using this functionality, they will likely use the test failover process in SRM to validate it before an actual disaster; however, the SRDF SRA does not support setting the parameter AutoTargetDeviceReuse=YES in this scenario. It must be set to “NO”. This is because of the SRM Cleanup workflow.
When AutoTargetDeviceReuse=YES and the user executes an SRM Cleanup, the SRDF SRA will determine the existing linked targets for the snapshots and then update the EmcSrdfSraTestFailoverConfig.xml file with these pairings. This ensures that when the user executes the next test, the SRA uses the existing devices for the test rather than create new ones. When two or more recovery plans are run simultaneously, however, the SRDF SRA may not be able to update the XML file for one of the plans while it is being edited for the other plan. This results in partial success and will cause the SRA to create new targets for one of the plans during the next test failover. Now depending on how SRM executes the plans (impacted by CPU, memory, etc.), it is possible that the EmcSrdfSraTestFailoverConfig.xml is updated properly with both device pairings from both plans; but it is uncontrollable. Therefore, the SRDF SRA will not support AutoTargetDeviceReuse=YES with concurrent recovery plans. It is important to ensure that before running simultaneous plans that any previous test failover runs are cleaned up using AutoTargetDeviceReuse=NO.
Note: In the event the user fails to adhere to this restriction and executes and cleans up concurrent recovery plans with the parameter enabled, manual cleanup will be necessary. This will involve removing the target devices from the recovery site storage group, removing the targets from the device/composite group, and unlinking and terminating any snapshots.