patternsqlMinor
SQL Server - memory issue after upgrade to 2022 CU7
Viewed 0 times
after2022sqlissuecu7memoryserverupgrade
Problem
In one of the SQL Server instances I am not able to execute any query. I thought it's due to some heavy load so I try to execute
]
Last night, the DBA has performed upgrade to SQL Server 2022 CU5 and told me there is an issue, that is resolved with CU7. So, upgrade to CU7 was performed, but nothing changed.
The DBA send me this:
and after that this:
but I am not sure what's going on. The DBA stopped the query store and all extended event sessions but nothing changed. Now, the DBA thinks the issue is caused by this but it should be fixed in CU4 and it causes problem during the upgraded, not after it.
The DBA has opened ticket to MS, but I am wondering if someone knows what may be causing this?
I am going to attach the trace data. A little more info - we have SQL Server 2022 CU5 instance and another one running the replicas with CU7.
The other DBA believes the issue is with the always on replicas, so they are stopped - the replicas were constantly crashing.
sp_execeute whoisactive and get:]
Last night, the DBA has performed upgrade to SQL Server 2022 CU5 and told me there is an issue, that is resolved with CU7. So, upgrade to CU7 was performed, but nothing changed.
The DBA send me this:
and after that this:
but I am not sure what's going on. The DBA stopped the query store and all extended event sessions but nothing changed. Now, the DBA thinks the issue is caused by this but it should be fixed in CU4 and it causes problem during the upgraded, not after it.
The DBA has opened ticket to MS, but I am wondering if someone knows what may be causing this?
I am going to attach the trace data. A little more info - we have SQL Server 2022 CU5 instance and another one running the replicas with CU7.
The other DBA believes the issue is with the always on replicas, so they are stopped - the replicas were constantly crashing.
Solution
Based on the logs that were uploaded, the errorlog has the first issue as Dump0062 (which is concerning by itself), anything that happens after that is suspect of being an actual issue as it could be a side effect of this issue leaving SQL Server in an inconsistent state.
Having said that, the initial issue looks to be with UCS:
This happens to be an
Based on the logs (and perceived configuration and setup from them), the error, the most likely cause is either a race condition or non-thread-safe access to data structures. If this is only happening on a single server (and all others are patched) then it may make sense to run a memory check (just to be safe). If it's happening on all servers, it's most likely a product issue.
If this wasn't happening on downlevel patches then it may make sense to uninstall the CUs that were installed and see if it reoccurs.
Having said that, the initial issue looks to be with UCS:
(Inline Function) sqllang!ListBase::Append
(Inline Function) sqllang!SEList::Append
sqllang!CSbPriorityScheduler::Append
(Inline Function) sqllang!CSbDestinationManager::NotifyDestinationStatusChange
sqllang!CSbDestinationManager::CDestination::HandleEvent
sqllang!CSbDestinationManager::CDestination::Subscribe
sqllang!CSbDestinationManager::SubscribeToDestination
sqllang!CSbServiceManager::SubscribeSession
sqllang!CSbServiceManager::RegisterSession
sqllang!CSbTransmitter::RegisterSession
sqllang!CSbMsgDispatcher::RegisterWithTransmitter
sqllang!CSbTransmission::RegisterWithTransmitter
(Inline Function) sqllang!CSbDeliverySM::SendMessages
sqllang!CSbDeliverySM::HandleEvent
sqllang!CSbTransmission::IdempotentWork
sqllang!CSbTransmission::IdempotentInvoke
sqllang!CSbTransmission_Task::IdempotentInvoke
sqllang!CSbTaskManager::HandleReusableTask
sqllang!CSbTask::TaskStart
sqllang!CSbTask::BeginTaskStart
SqlDK!SOS_Task::Param::Execute
SqlDK!SOS_Scheduler::RunTask
SqlDK!SOS_Scheduler::ProcessTasks
SqlDK!Worker::EntryPoint
SqlDK!ThreadScheduler::RunWorkerThis happens to be an
EXCEPTION_ACCESS_VIOLATION otherwise known as an AV. This causes SQL Server to shutdown (SQL Server is terminating this process.), so anything after this is nothing that needs to be investigated.Based on the logs (and perceived configuration and setup from them), the error, the most likely cause is either a race condition or non-thread-safe access to data structures. If this is only happening on a single server (and all others are patched) then it may make sense to run a memory check (just to be safe). If it's happening on all servers, it's most likely a product issue.
If this wasn't happening on downlevel patches then it may make sense to uninstall the CUs that were installed and see if it reoccurs.
Code Snippets
(Inline Function) sqllang!ListBase::Append
(Inline Function) sqllang!SEList<CSbPrioritySchedulableUnit,0>::Append
sqllang!CSbPriorityScheduler::Append
(Inline Function) sqllang!CSbDestinationManager::NotifyDestinationStatusChange
sqllang!CSbDestinationManager::CDestination::HandleEvent
sqllang!CSbDestinationManager::CDestination::Subscribe
sqllang!CSbDestinationManager::SubscribeToDestination
sqllang!CSbServiceManager::SubscribeSession
sqllang!CSbServiceManager::RegisterSession
sqllang!CSbTransmitter::RegisterSession
sqllang!CSbMsgDispatcher::RegisterWithTransmitter
sqllang!CSbTransmission::RegisterWithTransmitter
(Inline Function) sqllang!CSbDeliverySM::SendMessages
sqllang!CSbDeliverySM::HandleEvent
sqllang!CSbTransmission::IdempotentWork
sqllang!CSbTransmission::IdempotentInvoke
sqllang!CSbTransmission_Task::IdempotentInvoke
sqllang!CSbTaskManager::HandleReusableTask
sqllang!CSbTask::TaskStart
sqllang!CSbTask::BeginTaskStart
SqlDK!SOS_Task::Param::Execute
SqlDK!SOS_Scheduler::RunTask
SqlDK!SOS_Scheduler::ProcessTasks
SqlDK!Worker::EntryPoint
SqlDK!ThreadScheduler::RunWorkerContext
StackExchange Database Administrators Q#330826, answer score: 6
Revisions (0)
No revisions yet.