We have the following configuration:
- ESXi 5.1.0 running on three clustered Dell 710 servers
- vCenter Server 5.1.0
- The clustered servers have a total of 88G of RAM
- Looked through the logs on the SQL Server and found the following:
- In the System log, Resource-Exhaustion-Detected appeared at the same time but the highest memory consumption was services.exe at 125MB. Doesn’t seem to be the culprit.
- In the Applications log, started getting MSSQLSERVER errors indicating ‘The Server was unable to load the SSL provider needed to login.’ These errors persisted through the downtime.
- In the SQL Server Log File Viewer, we see the following errors, though I’m not sure that they are at all related:
- A significant part of sql server process memory has been paged out. This may result in a performance degradation. Duration: 0 seconds. Working set (KB): 4464, committed (KB) 66976, memory utilization: 6%
- (About 4.5 hours later) Warning: the available page file has dropped below 4Mb. It is at this point – seemingly – that the memory usage has jumped up to near 100%.
SQL Server 2005 x86 running on a virtual Server 2008 R2 x64. All databases run fine for several days with memory usage seeming at ~25%. Suddenly, the memory consumption shoots up to ~98% and we have to reboot the server. This is based on information from the task manager and the measurements in vSphere Client. I understand that the task manager is not an accurate measurement when AWE is active but that’s the best I’ve got at the moment.
Here's what we've done thus far:
The virtual server has 8G of RAM allocated. However, within the vSphere Client, the Memory Granted and the Memory Consumed seem to be at the same value…meaning that the consumed value seems to be at 100% under normal operating conditions. Not sure why that is.
If anyone has any recommendations or areas to look, we would certainly appreciate your assistance! Thanks in advance.
Tony