Unhandled Exception errors and timeouts on application HELP!!

[JoJoGabor]

Have you sat down with the DBA and run a SQL trace to see what is happening. Did anything change 2 weeks ago? It could be a network issue as well as a server issue. The trace should show you where the timeout is occurring

[TheFaQQer]

No amount of resources is going to stop an unhandled exception - code it properly to handle all exceptions.

More resources MIGHT handle the timeout, it all depends on the underlying cause of the timeout - e.g. if the timeout is caused because of another user locking a record, then no amount of resources is going to make the timeout go away.

[TheFaQQer]

I seem to be stuck with the server team laying all the blame at the feet of the developers and their code.

Quite right, too.

[lilelvis2000]

You need to run Perfmon on the SQL Server and determine what's going on first. Check to see how many SQL parses (hard parses) it is doing and how much disk activity and look for excessive numbers of waits and blocking etc.. Determine if the application is:
memory bound, IO bound or CPU bound.

Memory bound - adjust accordingly
IO bound - shift files to faster devices
CPU bould - either recode or install on faster machine

Good luck!

[original PM]

Hi All,

Again thanks for the replies

With regards to whether something happened 2 weeks ago - apparently yes another databse was migrated onto the same server as ours and some of our resources were 'stolen' this has been resolved.

The DBA has been looking at it - apparently he made some tweaks which should speed up.

I will also ask the techy teams to look at the other areas you mentioned to ensure we can nail the reliability of the system.

Thanks again!

[lightng]

Some good suggestions so far for the little info we have. App, DB, or network - it could be any one or a combination thats causing the problem.

You could also ask devs to isolate the query / stored proc thats running and put through query analyser to look at execution plans. Depending on the outcome, it maybe that your techies can improve things through indexing, updating statistics or rewriting the query involved.

You can try a short-term fix by increasing the watermark (and this is not ideal really). Increase the timeout in the sql connection string or increase the command-timeout; again not enough info to suggest which. The latter will require a small code change and the former probably just a config change (depending on how the app is written).

If your devs, dbas, sysadmins are palming the responsibility off on each other. Get them to work together. Lock them all in a room and don't let them out until they find a solution.

[lightng]

Glad you found your solution.

[QUOTE=original PM;759987]With regards to whether something happened 2 weeks ago - apparently yes another databse was migrated onto the same server as ours and some of our resources were 'stolen' this has been resolved.[QUOTE]

Ahh, the benefits of change control

[original PM]

Ahh, the benefits of change control

yes - up to about 6 months ago the developer and the server administrator sat pretty much next to each other.

unfortunately in the name of progress the server admin bod was moved up to Manchester (from Stoke) and then a load of red tape thrown around him (in the name of some form of crazy It service managment solution) - so we went from having something that worked to something that didn't!

in addition to this we have a load of b*ll*ck jugglers playing politics so if you speak to someone who is not their friend nothing gets done (even if you go through the correct but every changing lengthy and pointless procedures)

anyway it was my first brush with the new server control team and I think they now have the message of what we need and how to support us going forward so I do not foresee any further issues of this nature.

[lightng]

yes - up to about 6 months ago the developer and the server administrator sat pretty much next to each other.

unfortunately in the name of progress the server admin bod was moved up to Manchester (from Stoke) and then a load of red tape thrown around him (in the name of some form of crazy It service managment solution) - so we went from having something that worked to something that didn't!

in addition to this we have a load of b*ll*ck jugglers playing politics so if you speak to someone who is not their friend nothing gets done (even if you go through the correct but every changing lengthy and pointless procedures)

anyway it was my first brush with the new server control team and I think they now have the message of what we need and how to support us going forward so I do not foresee any further issues of this nature.

Sounds all too familiar. If you'd have said Bangalore instead of Manchester, I'd have sworn we were both working for the same client.

[MPwannadecentincome]

yes - up to about 6 months ago the developer and the server administrator sat pretty much next to each other.

unfortunately in the name of progress the server admin bod was moved up to Manchester (from Stoke) and then a load of red tape thrown around him (in the name of some form of crazy It service managment solution) - so we went from having something that worked to something that didn't!

in addition to this we have a load of b*ll*ck jugglers playing politics so if you speak to someone who is not their friend nothing gets done (even if you go through the correct but every changing lengthy and pointless procedures)

anyway it was my first brush with the new server control team and I think they now have the message of what we need and how to support us going forward so I do not foresee any further issues of this nature.

oh well at least you get paid for it!