Recall problems

Description

Device problem is reported after a failed recall. The error message logged in the ivd-hsm_ProcessID.log file is the following: device busy

Explanation

See Workaround.

Workaround

Try to identify the problem as follows:
• Determine the status of the corresponding OpenArchive (OA) partition with the command:
ivdpartition --status PartitionName
• Check if the ivd-hsm process is running on the corresponding OA client.

Use the ps -ef | grep ivd-hsm command for this purpose.

Description

Recall job does not search for alternative copies if the OA medium is marked 'unusable' after it has been allocated for recall.

Explanation

This may happen in the following situation:
• A running OA recall job has determined from which OA medium the data will be recalled but is still waiting for that medium to become available (probably because another OA job has allocated the same medium).
• During this waiting time, the medium status is set to unusable but the recall job that is waiting is not notified of this change.
The recall job will continue to wait for the medium to become “good” (usable) again, and therefore it will not search for available redundant copies migrated to other OA media pools, if any exist.

Workaround

Using the ivdjob command, identify and terminate all OA recall jobs that are waiting for the unusable OA medium to become “good” (usable). If there are redundant copies of the migrated data in other OA media pools, the subsequent OA recall jobs will allocate an OA medium from one of those pools, and recall the data from it.

Description

Recall from an OA medium fails and writes the following error to the OA event log:
MAJOR INTERNAL ERROR: Data corruption detected. 'State on medium doesn't match FSC. See error.log.

Explanation

Migration of the requested offline file generation has not completed successfully, most probably because the physical end-of-medium notification was not reported by the OA drive soon enough, resulting in part of the file data not being written to the tape.

Workaround

None

 

Description

When you modify attributes of an offline file through a Samba share, the file is recalled from the OA media.

Explanation

The customized Samba server running on a Linux system apparently reads part of the file contents after a file attribute is changed remotely from a Samba client.

Workaround

None