Seems I've hit on one that no one has any feedback on. I do seem to get those.
I'll be punting this to Veritas early next week - PTO starts in about an hour and a half, and I'd rather not be late for that!
I did do the same trace on the diff. backup and yup. it does not generate a SC 90 at any point. To be honest, It would be better if the job simply didn't fire up, but I do get how product works most of it's logic.
I'm a little confused - what exactly is in the backup selection for the policy - can you share please?
From the detailed job status it appears that the client is being told to backup X:\pagefile.sys
If it is included then the behaviour is expected.
Whups - meant to put that in, but forgot. The file list is X:\*, with multi streaming enabled, so the parent is passing the sub job the path of X:\pagefile.sys. As I said, the full backup generates the error, but the incremental doesn't for it's child job.
So, the environment is now at 8.2 for all but two clients, and as this is the media server, it's not a concern. The problem still persists. Looking back in history, it started about 2 months ago. Prior to that, the pagefile.sys is not getting listed as a problem child job. The Monthly job runs all child jobs with the correct schedule, and the Weekly job has gone back to running the pagefile.sys job as a Monthly schedule, when all the others are Weekly (I'm talking about the User friendly schedule name, both are full's - just having different retention levels). The incremental job starts a child for pagefile.sys, sees it's on the exclude lists, and exits gracefully with SC 0.
So davidmoline, I don't think this is a "by design" element. If the incremental can cleanly detect it should not backup an item in the exclude list, and exit gracefully, a full backup should as well. Clearly something changed 2 months or so ago, but I'll be darned if I can tell you what. There were the usual OS patches applied but that would have been 2+ weeks prior. The data itself changes on a regular basis - it is MS SQL backups (dump and sweep style vs. using the agent), no patches to the NB binaries occured prior to Dec. 3. This is a very calm, conservative NB environment - nothing wild or cutting edge is being done.
Since the problem exists in NB 8.2, I'm taking it over to Support this afternoon.
I assume that the X: volume is otherwise empty, so what happens if you create a placeholder file, e.g. named: do-not-delete-me.txt Does the full backup still fail with status 90 for that volume ?
