Hello all,

We had a helpdesk frontliner trigger an interesting...phenomenon...in Exchange 
yesterday.  I don't have the full details, but I was told he was trying to copy 
a calendar folder from his Live@edu mailbox into his on-premise Exchange 
mailbox (Outlook 2010; he had both mailboxes in separate Exchange-type 
accounts).  For whatever reason, the copy failed, and he went to do something 
else.  When he went back to Outlook a minute or two later, instead of the 
process actually "failing" he now had 232 (I counted) Calendar-type folders.  
These folders were in one single hierarchy, like so:

/Calendar
/Calendar/Work
/Calendar/Work/Calendar
/Calendar/Work/Calendar/Work

And so on, until the final "Calendar" folder was 232 levels deep, with a path 
length of 1,622 characters.  The "Calendar" folders had zero items in them, 
while I think the "Work" folders had maybe six items each.
Anyway, said student tried to delete the folder structure in Outlook but 
couldn't.  (I don't know what error it gave him there.)  He then tried to 
delete the structure in OWA, but it didn't work there, either; it kept giving a 
"permission denied" message.  (I theorize that OWA didn't know *why* it 
couldn't delete the folder, so it defaulted to "you must not have permission".  
But I could be way off-base.)

So, they called me.  I loaded up the mailbox in Outlook, saw it fail, so I went 
to MFCMapi and tried to delete it there.  However, every time I tried, the 
mailbox would disappear for a few seconds, come back with a 
MAPI_PARTIAL_FAILURE (that's not quite correct, but I didn't write it down 
yesterday), and not actually delete anything.  I then decided to export his 
mailbox as a PST and exclude the /Calendar/Work/* folders in order to reimport 
into a "clean" mailbox.  As I'm running "Get-MailboxExportRequestStatistics", 
in a slow loop, the cmdlet errors out with an error similar to "Cannot get 
statistics for request because the database is dismounted on the mailbox 
server."  Uhh...what?

To make a long story endless, it turns out that something to do with this huge 
folder structure seems to have been crashing the information store, or at least 
caused it to dismount the database...and then all the rest of the databases on 
that server seemed to dismount as well.  The "crash" happened multiple times 
before I killed the export request, and I went back and found out that every 
time I tried to delete one of the folders using MFCMapi it would also cause the 
database to dismount-hence why the mailbox "went away" for a few seconds, until 
the DB was remounted on another DAG member.  It was at about this point that 
the frontliner called back and told me it was okay to just delete his mailbox 
and restore from backup, which was A-OK by me.

So my question to y'all is:  has anyone seen that behavior before? Is there a 
"max folder depth" in Exchange 2010?  (I couldn't find one via Google.)  A max 
folder path character count? (Again, I couldn't find one using Google.)  And 
even if there was, is the expected result that "...the database <blah> was 
stopped" (MSExchangeIS Mailbox Store 9539)?

Thanks for any help or cluesticks anyone can provide.

Seth Wright
James Madison University

---
To manage subscriptions click here: 
http://lyris.sunbelt-software.com/read/my_forums/
or send an email to [email protected]
with the body: unsubscribe exchangelist

Reply via email to