Disk volumes that are under LifeKeeper protection do not have automatic boot-time CHKDSK performed on them. Administrators should occasionally perform CHKDSK on their volumes manually. To do this, follow these instructions:
- Take all resources that depend on the volume out of service. During CHKDSK, the volume will be locked and inaccessible, so any applications that are using the volume should be taken out of service.
- Leave the volume resource in service – CHKDSK needs to be able to open the volume, so it should be left in-service and unlocked.
- If the volume is a DataKeeper mirrored volume, we recommend that the mirror be put into a Paused state.
- Put the volume resource in maintenance mode by running these commands from the command line on the node where the volume is being checked:
- cd %LKBIN%
- flg_create -f Maintenance_<drvletter>
- <drvletter> is the volume drive letter – to put volume F: in maintenance mode, run the command: flg_create -f Maintenance_F
While the volume is in maintenance mode, volume check scripts will not perform health checks.
- Check the volume and fix errors (for example, run “chkdsk.exe /f F:“).
When CHKDSK completes:
- Take the volume resource out of maintenance mode by running these commands from the command line on the node where the chkdsk completed.
- cd %LKBIN%
- flg_remove -f Maintenance_<drvletter>
- <drvletter> is the volume drive letter – to take volume F: out of maintenance mode, run the command: flg_remove -f Maintenance_F
- Continue any mirrors that were paused to facilitate performing chkdsk. This allows filesystem changes to be resync’ed with target nodes.
- Bring hierarchies in service. These can be brought online while the mirror is still resyncing.