Testing a StorPool drive

This procedure describes how to manually test a StorPool drive marked for testing by StorPool as per the disk-to-test monitoring alarm.

1. Put node in maintenance

Set the server on which the disk is physically mounted in maintenance mode with:

# storpool maintenance set node <nodeId> duration 15m description testing_disks

2. Perform test

2.1. Check disk’s server instance

Check if the disk that needs to be tested is the only one in that server instance.

# storpool_initdisk -l

/dev/sda2, diskId 111, version 10009, server instance 0, cluster xxxx.x, WBC
/dev/sdb1, diskId 112, version 10009, server instance 0, cluster xxxx.x, WBC
/dev/sdc1, diskId 113, version 10009, server instance 0, cluster xxxx.x, WBC
/dev/sdd1, diskId 114, version 10009, server instance 1, cluster xxxx.x, WBC

This is an example of a single disk /dev/sdd1, diskId 114 in server instance 1.

2.2. No single drives on server instances

If there is more than one disk in that storpool_server instance, you can trigger the disk testing using the following command:

# storpool disk <diskID> eject test

2.3. Single drive on server instance

To invoke a testing procedure in case of a single disk in a server instance, follow the steps below (example for disk 114(/dev/sdd1) running on server instance 1):

  1. You need to stop its storpool_server instance first.

# systemctl stop storpool_server_1.service
  1. Mark the disk for testing.

# storpool_initdisk -r --test-override test 114 /dev/sdd1
  1. Start the storpool_server instance again.

# systemctl start storpool_server_1.service

3. Validate result

3.1. Disk comes back

In both cases, if the disk comes back into the cluster after the test, that means it passed the testing. You can now complete the maintenance mode with:

# storpool maintenance complete node <nodeId>

3.2. Disk doesn’t come back

In case the disk does not come back into the cluster, verify the reason in the logs. If the drive doesn’t pass the tests, move on to removing the drive from the cluster, or if you have spare drives, proceed as described in 5.  Removal & re-balance out.