Monitoring Stopped

eilandd

Since upgrading to version 2.2 we have received a rash of Monitoring Stopped alerts. We did not have any monitoring issues prior to upgrading.
These alerts are not generated immediately but after a period of successful monitoring. Monitoring is restartable but will fail after a few minutes. All alerts are showing "Incorrect credentials or insufficient permissions". We have tried several different credentials, all with administrative access, with no change in behavior. Any help would be appreciated

Chris Spencer

Have you checked the Show Log page just after receiving a Monitoring Stopped alert? This should give some more details about what is actually causing monitoring to stop. There is a Show Log link for each server or instance on the Monitored Servers page. The log only displays the last 5 minutes so you have to be pretty quick but hopefully there will be a more specific description of the error.

Regards
Chris

eilandd

Yes we have looked at the log. The latest error is

GetBinaryValue: \\prod-db4\HKEY_PERFORMANCE_DATA\236 510 Incorrect credentials or insufficient permissions Win32Exception Access is denied


This issue started after the upgrade to 2.2. Target servers are Windows 2008 R2.

Chris Spencer

It might be a good idea to test the remote registry connection to that server using the steps documented on this page:
http://www.red-gate.com/supportcenter/C ... 402836.htm

The monitoring alerts were introduced in version 2.2 so it's possible that there were intermittent errors previously but you were not informed of them.

Hope this helps
Chris

eilandd

All manual tests pass.

Chris Spencer

Are you only getting the monitoring stopped alerts for a single server or are they all having this issue?

Regards
Chris

eilandd

At this point there are 2 that are experiencing this issue.
Both are physical Windows 2008 R2 servers in different domains from the base monitor.

Chris Spencer

We didn't as far as I know change anything in the 2.2 release that would affect data collection in this way. I'll do some specific testing with Windows 2008 R2 to see if I can reproduce the issue.

Quick question, the account that you use to connect to these 2 servers, is it the account running the base monitor service or do you specify a different one?

Regards
Chris

eilandd

Since these servers are in a different domain than the base monitor we use a login in the target domain.

To clarify what we're seeing. The issue is not affecting servers in the same domain as the base monitor. The servers that are being affected will successfully work for a while before failing. The failing servers are all Windows 2008 R2. Sometimes the failures are at the host level and sometimes at the SQL instance level. The failures are always Access Denied on the Performance registry hive. This morning we had the failure on the passive node of an active\passive cluster while the active node had no issues. Currently, one the servers that has been an issue is working while the other is failing at the instance level but reporting normal at the host level. We have tried both windows and sql credentials and still see the issues. Here's the current failure, this is with the same windows login that is working at the host level:

GetBinaryValue: \\dev-db\HKEY_PERFORMANCE_DATA\15886 15666 15796 15734 15870 15778 16034 15974 Incorrect credentials or insufficient permissions Win32Exception Access is denied

Chris Spencer

Thanks for the extra information. I've raised a bug report (reference SRP-3585) to investigate this.