Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I'm trying to diagnose what is causing regular crashes and hangs in one of our database applications. Operating system is windows server 2003, database server is SQL server enterprise 2005.

We recieve a large number of errors like this prior to the crash/hang.

Error:  OLEDB returned result 80004005. Description: E_FAIL.  Unspecified error

I've spoken to the software vendor and they believe this is being caused either by either excessive load on the SQL server or by a loss of connectivity with the SQL server.

I do not believe this is a load issue - if anything it appears worse overnight when there is less load on the SQL server (it's possibly an idling issue?). Connectivity is a possibility however other databases on the server do not experience issues, and another application accessing the database is not reporting any issues (although it may be more fault tolerant).

Running tests on our test system I can occasionally replicate this issue by blocking the SQL server port via windows firewall for ~20 seconds and then reenabling - sometimes it results in the same behaviour, sometimes the application recovers. I've written a utility to try and connect to the database every second to check for connectivity, however it didn't show any loss of connectivity.

What next steps can I take to prove/disprove the SQL server connectivity is good?


share|improve this question
"Excessive load" and "loss of connectivity" are about the 2 broadest finger-pointing cop-outs ever... E_FAIL is a pretty broad error description, though. – squillman Sep 13 '10 at 13:50
i have had a very similar conversation with the software vendor. ;) – Rob Rob Sep 13 '10 at 14:01
Anything useful in the SQL logs or the windows event logs? – doug_w Sep 13 '10 at 15:59
the sql logs are very quiet. They detail backup successes and how long the SQL server has been using the process ID for. – Rob Rob Sep 13 '10 at 16:26
the system event log just has notes in. – Rob Rob Sep 13 '10 at 16:32
up vote 0 down vote accepted

It looks like this issue was being caused by tcp offloading, which has known issues with the Broadcom 5708 chipset:

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.