We have 48GB of ram. Thank you for the link, but it does not seem to apply. Is there some reason why you are thinking this will help? We are not really having cluster issues.
Well, I recommend that to help alleviate any race conditions. It should be applied regardless.
Having said that, 4GB of RAM seems awful low. How many users? Can you double the RAM to 8GB or even 16GB?
How much RAM in the server? Also, to ensure you are at the latest, appy SP2 RU$.
Finally, Apply the hotfix described here:
Recommended Windows Hotfix for Database Availability Groups running Windows Server 2008 R2
Hi,
Please use the performance monitor and compare with the performance counters to find the root cause:
Perfwiz for Exchange 2010
http://blogs.technet.com/b/mikelag/archive/2010/01/11/perfwiz-for-exchange-2010.aspx
Performance and Scalability Counters and Thresholds
http://technet.microsoft.com/en-us/library/dd335215.aspx
Thanks,
Simon
We have an antivirus solution installed, but we have set exclusions. We have also tried disabling it and the issue still occurs. We do use snapmanager for exchange.
No drops in connection with the cluster.
I have not run the ExPBA, I will do that and report back.
If we restart the service or the server, the processor utilization returns to normal, but then starts to climb again. I would say with our 8 vcpus, we get about 2 and a half weeks before the issue forces us to restart the server or service.
can you check what processes are taking your cpu
Hi
Do you have any 3rd party products installed?
Do you have file level anti-virus scans enabled?
Do you have any drops in connection with your cluster?
Is anything logged in event viewer?
Can you run ExBPA and check if it generates any issue/warning?
If you restart the service does the utilization drop?
Hello,
We have an exchange server 2010 sp2 installation that is about 6 months old. We have always had an increasing processor utilization issue since the environment was built that we have not been able to resolve. Our setup is as follows:
We have a DAG with two member servers, and one witness server. The servers are located in separate data centers with a 100Mbps link and <1ms latency. There is only one AD site for both data centers, there is not an AD site per datacenter. I believe the subnet/nic configuration is different than what Microsoft recommends. We have 2 nics per server, replication and local area connection. On one server, both nics are on a .0 subnet, and on the other server, both NICs are on a .47 subnet.
I have a theory that the lack of seperate AD sites, and the non-standard subnet configuration could be contributing to, if not causing our replication issues. Is this the wrong line of thinking? Thanks!
Hi,
Did you succeed in solving this issue? We have exactly the same problem. after running 2 weeks CPU usage is growing where msexchangerepl.exe is taking more CPU usage by every day. After running 4 weeks with the server the total CPU usage is 100% of the server.
We are running in an exchange 2010 SP3 DAG with 2 mailbox servers with 4 mailbox databases serving 1800 mailboxes and running 2 hubtransport/client access servers.
best regards,
Emiel
Hi,
Did you succeed in solving this issue? We have exactly the same problem. after running 2 weeks CPU usage is growing where msexchangerepl.exe is taking more CPU usage by every day. After running 4 weeks with the server the total CPU usage is 100% of the server.
We are running in an exchange 2010 SP3 DAG with 2 mailbox servers with 4 mailbox databases serving 1800 mailboxes and running 2 hubtransport/client access servers.
best regards,
Emiel
Yes, both of our servers have Hub transport, Client Access, and Mailbox roles installed. Great link thank you.
The problem continues with the increasing CPU usage on the primary server. The graph looks the same as above.
What can I check? Thank you!!
When we ran the exbpa, it complained about disk latency on our passive server. We moved that server and disks to a new san which took care of that.
We also fixed the networks so that they match what Microsoft recommends.
Finally, we installed the hotfixes today. I will update with a graph soon.
Question though:
our passive server has mail flowing through it according to headers and our spam appliance. i would think we should only have mail flowing through it when the dbs are active. Is this configured with the CAS array? could the CAS configuration be related?
So you have the hub role installed with the MBX role?
See for an explanation of behavior if so
http://technet.microsoft.com/en-us/library/ff699049.aspx
Hub Transport and Mailbox Server Roles Coexistence When Using DAGs
When we ran the exbpa, it complained about disk latency on our passive server. We moved that server and disks to a new san which took care of that.
We also fixed the networks so that they match what Microsoft recommends.
Finally, we installed the hotfixes today. I will update with a graph soon.
Question though:
our passive server has mail flowing through it according to headers and our spam appliance. i would think we should only have mail flowing through it when the dbs are active. Is this configured with the CAS array? could the CAS configuration be related?
thank you for the explanation. We have Fourty Eight Gigabytes of Ram, so I think that should be enough. I am going to apply these hotfixes to see if it helps. Thanks again.
Ah ok, Sorry. Misread that! lol. I think I need glasses! :)
We have 48GB of ram. Thank you for the link, but it does not seem to apply. Is there some reason why you are thinking this will help? We are not really having cluster issues.
Well, I recommend that to help alleviate any race conditions. It should be applied regardless.
Having said that, 4GB of RAM seems awful low. How many users? Can you double the RAM to 8GB or even 16GB?
How much RAM in the server? Also, to ensure you are at the latest, appy SP2 RU$.
Finally, Apply the hotfix described here:
Recommended Windows Hotfix for Database Availability Groups running Windows Server 2008 R2