Last week i was monitoring the SAN storage with VNX Monitoring and Reporting. I saw around 7000 IOPS coming from one LUN. I changed the period to 3 months and saw from on one day the IOPS raise from around 200 IOPS to 7000 IOPS! WOW!! Looked in to VMWare vCenter and nearowed it down to the SCCM primary site server. Started performance monitor and most IOPS did come from the partitition where SCCM is installed. After that i started PROCMON on the SCCM server and used file summary to analyse what did cause the IOPS.
A lot of file events or coming from \inboxes\schedule.box\outboxes\LAN folder. I checked the \inboxes\schedule.box\ and there are a lot of .JOB files. In the \outboxes\LAN folder contained a lot of SRQ and SRS files. After that i checked the log files Sender.log and schedule.log and found a lot of information about intersite replication occuring from the primary site to the secondary sites. Most secondary sites are offline but there were a lot of jobs for the secondary sites. I had the feeling this was messing up my primary site and did cause a lot of IOPS. Did some searches on google and the following was the resolution.
1. Stop all SCCM services on the SCCM server
2. Start CMD as Admin and go to <driveletter>:\<program files>\Microsoft Configuration Manager\bin\i3860000409
3. Delete jobs with the following command. Preinst.exe /deljob <sitecode>.( Sidecode is the site that is offline and causing the issues. after deleting the site 2700 jobs were deleted!)
4. Deleted all old SRQ and SRS files from the \outboxes\LAN and Requests folder.
5. Start all SCCM services.
After this IOPS went fromt 7000 IPS to 200 IOPS! This is the expected behaviour! 🙂