Comments
Sort by recent activity
I managed to get the hotfix rolled out on one server yesterday morning, and haven't had a connection failure since. That's a success in my book [image]
I'll be rolling that hotfix out to all Windows 2008/r2 servers over the next month to make eliminate the rest of the connection failures I get. / comments
I managed to get the hotfix rolled out on one server yesterday morning, and haven't had a connection failure since. That's a success in my book
I'll be rolling that hotfix out to all Windows 2008/...
The best idea I've come up with so far is a pair of memory leaks in Windows - particularly one related to WMI. After a while, the memory allocated to the wmiprvse.exe service will reach 512MB, which is a cap - at this point any remote WMI calls will fail. After a few seconds, some garbage collection will occur to free some memory, and SQL Monitor will connect in again.
I've scheduled a hotfix to be applied, but my OS team is slow to roll these things out, so i can't state if this will definitely solve the problem...
For ref, the Windows 2008r2 hotfix is here: http://support.microsoft.com/kb/2832248, and the vanilla 2008 one is here:http://support.microsoft.com/kb/958124
If that sorts out your issues, please let me know [image]
Rich / comments
The best idea I've come up with so far is a pair of memory leaks in Windows - particularly one related to WMI. After a while, the memory allocated to the wmiprvse.exe service will reach 512MB, whic...
Out of interest, is the problem server a Win 2008 R2 box? I'm currently investigating various Machine Unreachable/Monitoring errors, and have found a rough correlation with Out of Memory exceptions in the Base Monitor log. These exceptions may be thrown by a misbehaving WMI provider service on the monitored box, in which a memory leak causes the service to increase in memory usage until it reaches a 512MB cap. The service will eventually recycle, but will effectively disable remote monitoring until it does.
I'm not saying that this is definitely my problem, or yours, but it is a potential line to check, particularly if (as in my case) you have multiple monitoring systems which both use WMI.
For ref: http://support.microsoft.com/kb/981314 / comments
Out of interest, is the problem server a Win 2008 R2 box? I'm currently investigating various Machine Unreachable/Monitoring errors, and have found a rough correlation with Out of Memory exceptions...
I checked the 'Show log', but never managed to get to it in time to see anything helpful - that log appears to only retain a few minutes of info...
As an experiment, I moved the Base Monitor to a different server, and I haven't had a connect error since the move. It's only been two hours, but I'll keep my eye on it and hope that it was a problem with the monitor host.
Things couldn't be totally fixed, of course - I've now got 100% CPU usage on the new host [image] I'll raise that in a new thread if it doesn't settle down this afternoon... / comments
I checked the 'Show log', but never managed to get to it in time to see anything helpful - that log appears to only retain a few minutes of info...
As an experiment, I moved the Base Monitor to a d...
Thanks for the suggestion Brian, but login failure auditing was already on and no SQL Monitor related events are in the SQL Server log. I have seen other user login failures, so is it definitely logging...
This is one of the odd things about the problem - SQL Monitor saying that it has had SQL credential problems, but the SQL Server itself denies all knowledge. Most peculiar. / comments
Thanks for the suggestion Brian, but login failure auditing was already on and no SQL Monitor related events are in the SQL Server log. I have seen other user login failures, so is it definitely lo...
Btw, I've also checked the SQL Monitor 'Monitored Servers' machine log, but that only goes back about 3 minutes.
The full SQL Monitor log shows all sorts of random exceptions, but none with times that correspond with the monitoring failures... / comments
Btw, I've also checked the SQL Monitor 'Monitored Servers' machine log, but that only goes back about 3 minutes.
The full SQL Monitor log shows all sorts of random exceptions, but none with times t...