I'm struggling with my understanding of the 'Delete Orphans' feature within CTA, and how best to apply a good 'Delete Orphans' policy.
We have CTA/VE version9.0 being used to archive content from VNX to Centera. All file data hosted on the VNX is subject to a backup policy with 6 months retention. We therefore want to guarantee that any deleted stubs can be restored under this policy and still connect to the associated Centera tier file content.
Reading the CTA documentation would suggest that a 'Delete Orphans' policy can be configured to keep Orphaned files (files with a deleted/missing stub) for a period to match the backup policy, and longer that the default 30 days. Therefore the CTA would protect the file content with missing stub from being deleted by Centera for the duration of the 'Delete Orphans' policy.
The two main points I am struggling with are;
I think you have understood the CTA Orphan Management very well.
For each Archiving sourcey that you define in a schedule, the CTA will automatically create a stubscanner schedule.
You can get a list of the stubscanner schedules by picking them in the schedules window of the CTA GUI and you modify them as well.
Once a stub gets delete on an archiving source and a stubscanner runs on this source, it will mark the object as "orphaned" in the CTA database with the current timestamp.
If you now create an 'delete_orphans' policy and set the "Days missing more than" parameter to 6 month, this orphan_delete job will only delete object from the secondary storage that have a orphaned timestamp in the database that is at least 6 month ago.
Sorry for the late reply - its taken a while to get these policies settled and tested. Still, it's given me chance to test orphan delete policy, and the recovery of stubs older than a couple of months. All appears to be working well. Thanks for you're comments.
When you created your inital archive policy on the CTA you would have been given the option to have retention or not. If you picked 6 months, then the filescan not be deleted until 6 months after being archived. I do not think this is what you are looking for.
To get what you want you need to determine the correct settings in the delete_orphan policy. You need to set the days missing to 7 months when creating your delete orphan policy. This will ensure that the files can be restored at least 6 months after being deleted. For the rule on the delete policy, just pick something like size < 10 TB to make sure you get all files.
Now why 7 months you ask? It is very confusing. It is 7 months since last seen by the Stub Scanner.
1st of Jan - Stub Scanner run on FS - file is seen
14th of Jan - Stub file deleted
1st of Feb - Stub Scanner run on FS - file not seen
2nd of July - Orphan Delete run - The file has a last seen of more than 6 months (last seen on Jan 1st) so file on target is deleted, but the file is only deleted 5 and a half months ago.