1

i have a SQl 2016 AlwaysOn cluster that has almost nightly issues. I've narrowed it down to it going down due to errors reported in sp_server_diagnostics.

This is a warning right before the cluster failed last nigh before there are health check timeouts and other messages saying always on is unhealthy

<system spinlockBackoffs="0" sickSpinlockType="none" 
 sickSpinlockTypeAfterAv="none" latchWarnings="2" 
 isAccessViolationOccurred="0" writeAccessViolationCount="0" 
 totalDumpRequests="2" intervalDumpRequests="0" nonYieldingTasksReported="0" 
 pageFaults="1882" systemCpuUtilization="35" sqlCpuUtilization="7" 
 BadPagesDetected="0" BadPagesFixed="0" LastBadPageAddress="0x0"/>

Is there a specific entry I should look at first?

1

1 Answer 1

1

Ensure your servers follow VMware's Best Practices Guide for SQL Server.

Specifically, ensure the VMs running SQL Server have 100% of their memory and CPU reserved in VMware. Ensure you're using the PVSCSI (para-virtualized SCSI) adapter for SQL Server RDMs or Virtual Disks. Try to use as many PVSCSI adapaters as you can since that will increase the number of disk queues available to SQL Server.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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