Had an issue today with a client where some of their users stopped getting new messages. The Information Store service was running and all databases were mounted.
After they contacted me I confirming IS and databases were on-line then I checked disk free space. The transaction log drive had fewer than 10% free space so I assumed it was a back pressure issue. The “back pressure” feature was introduced in Exchange 2007 and monitors critical system resources like memory and disk space, in relation to the Transport process. When resources are low the transport process will start limiting or stop message delivery. I’ve run into this in the past when building VMs, with very tight disk space, since the default is 500 MB or 2% of free space. So I disabled back pressure monitoring by setting Enable=False in the EdgeTrasport.exe.Config file. I then restarted the transport service but this did not fix the problem. See this article for more details: Understanding the Back Pressure Feature in Exchange Server 2007
The next thing I checked was what was taking up 90% of the space on the TLOG volume. It turns out there were transaction logs going back almost two months, which means the client’s Exchange aware backup solution wasn’t working or is non-existent (it turned out to be the later). So I moved the oldest TLOGs to another drive to free up disk space. I don’t delete TLOGs until I know that there has been a successfully backup of Exchange, which should delete the TLOGs for you. I then dismounted and remounted the database in question, probably not needed but I didn’t want to wait long to see if the freeing up of disk space fixed the issue. After a few moments users started getting their messages that were stuck in the queue.
What is odd about this issue is that I didn’t see warning or error events being logged to the Application log after restarting the Transport Service. After going back now and searching for just errors I did find the following events:
10014MSExchangeIS Mailbox Store
Available disk space for the database logs is extremely low. Message delivery to the database may be stopped.
Database: ‘User Mailbox Database 1510132252’.
The same event was logged for the MSExchangeIS Public Store also.
When I went into queue viewer there were 100s of messages in the queue for the database that continued the users who weren’t getting messages. In the Last Error column it has “431 4.3.1 storedrv mailbox disk is full”. I didn’t find anything on this error when searching, which is the main reason I’m writing this post. This was also the final clue on what was causing the issue.
Environment: Single Exchange 2010 server with two mailbox databases, each using their own DB and LOG LUNs.