We have Enterprise Vault 12 for Exchange 2010 and doing a massive clean-up of old mailbox archives with manual deletions from console. The Technote - https://www.veritas.com/support/en_US/article.TECH49804 has been referred for this exercise. Some of these archives have thousands of mails and understand that it may take time for complete deletion. After we have started deletions (say around 500 archives in a day). recently noticed that the archival is not getting completed in the schedule. My queries are
1) Will these deletions affect our regular archving schedule
2) Will there be any performance issues due to deletions
Appreciate your feedback.
Archive deletions have a (as you noticed) big impact on archiving. this is due to what happens behind the scenes. If I recall correct, when you delete an archive, first the index for the archive is emptied. then the items are being marked for deletion. then the items are deleted and the database is being updated. If you use CAB files, the impact is even bigger, as there is a count of how many items in a CAB file are marked for deletion. I believe the threshold is 15%. If 85% of the items in a cabfile are marked for deletion, the cabfile is extracted, and the files that need to be removed are removed. The other files are cabbed again.
In short, an archive deletion has huge impact on SQL and Storage. Seeing the amount of archive you are deleting in one go (500..) this might very well impact normal archiving. I would suggest you figure out which archives have little items, and process these first. Then do large archives, but only a few (I would start with 5 personally, but 10 or 50 might also be ok), and monitor. You should see lots of activity on SQL, and on storage.
I'm not really sure on the below, but support should be able to assist: I'm not sure what actually handles archive deletion (from the top of my head). If it is storage expiry proces, you might be able to schedule the deletions, to not interfere with the normal archiving.
My advise in short: start with a small amount of achives to delete, and then try to find a balance between archive deletion and archiving. You might also want to extend your archiving window (if that is possible), to make sure all your archiving is done in the run.
GertjanA's response is right on.
Deleting individual items (via EV Search, for example) is handled by the StorageOnlineOpns process. Administrative deletions (via right-clicking an archive or Vault Store in the VAC) are handled by the StorageDelete process. Neither operation can be scheduled like Expiry can. Rather, once the deletions are requested, they proceed until they are finished.
To determine which archives are largest and which are smallest, run the following SQL query against your Vault Store database(s).
ArchivePointId AS ArchiveID
,ArchivedItems AS ArchiveItemsCount
,ArchivedItemsSize AS ArchivedItemsSizeKB
,(ArchivedItemsSize / 1024) AS ArchivedItemsSizeMB
,(ArchivedItemsSize / 1024 / 1024) AS ArchivedItemsSizeGB
FROM ArchivePoint AS ap
JOIN EnterpriseVaultDirectory.dbo.Root AS r
ON r.VaultEntryId = ap.ArchivePointId
JOIN EnterpriseVaultDirectory.dbo.Archive AS a
ON a.RootIdentity = r.RootIdentity
ORDER BY ArchivedItems DESC
You can also take advantage of EV's reports and save the hassle of running SQL queries.
There's the 'Vault Store Usage by Archive' report in the 'Operation Reports' section, which can be exported to CSV and sorted in Excel. If you don't have SQL Reports set up, you can also go to http://<EVServer>/EnterpriseVault/usage.asp and save the report for that Vault Store by Archive. The resulting TSV file can then be imported into a spreadsheet for sorting as well.