I am running EV8 currently and I am receiving a couple of warnings in my event log.
The erros are 41008 (5) and 41021 (1)
I have researched the erros which has something to do with the savesets (safe copies) which get removed after a backup is completed. The thing I don't understand is all my partitions/Vault Groups have Remove Safety Copies set to Immediately after Archive.
I have read some things about using archive bits and a trigger file but I am not backing them up.
I am currently updating/reindexing all the mailboxes/journal folder to see if this helps. Is there a place I can manually run the monitor to see if the savesets are still there?
After I get this issue resolved hopefully a couple other problems will fix themselves.
I don't know those event id's from memory but my guess is they are ones reporting a backlog of entries in the watchfile table and then saying this could be related to a backup not having occurred.
So if you are not backing them up what are you doing to secure them, is this just a lab?
EV creates those watchfile records because it wants to ensure those items are safely backed up so that in can SIS with them later. Just because you have one partition for example set to immediatley after archive, it doesn't mean you may not be sharing it with another parititon set to 'after vault store backup' and also there could be sharing between these two if they are in the same vault store group.
So the mention of trigger file is a way for you to signal the partition is safely backed up which would then make those watchfile records go away.
I have all my partitions set to immediately after archive. This is a production server and I have not seen the benefit of backing up the safety copies yet. We have up to the minute snapshots of all mailboxes and we only archiving items a year old. The only difference of course being the Journal archive.
We use NetApp for everything and only use EV for Exchange purposes. Everything we have is in a snapshot/snapmirror.
Should this be different?
We use snapshots in Symantec as well. What we do is use a trigger file that gets generated as part of the snapshot. That way if the snapshots fail then we don't do the post processing.
By default EV will check for the trigger file every 1 hour or so but you can via a registry key change it to check for it more often (for example once every 5 minutes or so). So this should then give you what you want albeit a bit safer.
Don't suppose you know how I can do that? The CIFS snapshot we take is about every hour. Nothing is done by Symantec this is all done by NetApp. WE just have a CIFS share with the files in it and the index files are on local drive space.
One way you could do it is to schedule a task on the EV server to create the trigger file - have the task run every 12 hours or 24 hours.
You will need to create the trigger file (IgnoreArchiveBitTrigger.txt) in the root of the partition/s - then set the Partition to check for the existence of the file.
Hope that helps.