Skip to main content
commvault.com commvault.com

Hi guys,

due to  loss of a complete mailbox we had to restore all objects of the user back to an empty mailbox. Environment is EWS-based, hybrid Exchange, Mailbox v2, incremental forever...

The user has the habit to move all mails from inbox to folders, keeping inbox “clean”. Oftentimes this move happens only days after receiving the mail.

So after successfull restore the user found 80% of the mails in the inbox, because that’s where they were when the backup ran. While, yes, all objects are restored, it’s for from optimal behaviour, as you can imagine.

Apaprently moving mail objects from one folder to another in outlook means deleting the object at the source and creeating a new object (with a new GUID) at the destination. Commvault has no chance of tracking the object’s “journey” within the mailbox. 

So it’s “works as designed”. 

Since Full backups are not supported (which would probably index all objects as-is again), we’re stuck.

Is this something you have experienced / are aware of? Is there a known way to solve this?

Thanks,

Ingo

 

Anyone?

Apologies, @Ingo Maus , I’ve been checking with people since yesterday.  I’ll follow up.

Hello Mike,

thanks for your effort. We replicated the scenario in our test lab. Same thing. For what it’s worth, it may be interesting for other community members. We assumed the agent would mirror the mailbox content as an “image” if you will. It does not.

What’s worse is that once you restore “all objects” in bulk it will then restore quite a number of objects twice, which could result in more prod storage consumption: the first “unique item” back to the INBOX, the same object (now with new UUID) back to the folder it eventually went. 

 

I believe your experience is correct, and almost definitely a limit of the MSFT APIs (which is almost always the case).

I’m still asking around for some input, so hopefully I can get something for you to confirm.

Hi,

I’d like to ask again if there’s any way around that limitation. We’ve been discussing this and came up with the idea to somehow force the agent to reindex all mailboxes in regular intervals, like every month or so. That way at least we’d only lose 30 days of user’s “item organizing”.

MBv2 agent does not support Full backup, which would probably index all items. So we’re stuck.

I’m not sure if other solutions find a way around it. Would be interesting to know.

 

(edit): Let me add that there’s been a discussion here: 

There’s mentioning of “reindexing”. Would that “converted to full” job then do a full index and thus present the mailbox in the browse as it was at the time of that job? Any idea if that job takes considerably longer?

I’m tempted to try out.

 

 

Reply


Terms & ConditionsCookie settings