Hi all,
We have a small cluster of RDS boxes (2 for users) (on Hyper-V) to run MS Access for an in house application (connecting to SQL Server 2012), which has been running with live users (approx 50 per box) for about a month.
Each box has:
- burstable Ram upto 48GB (not gotten close to it),
- 4 x virtual cores,
- Windows Server 2012 R2,
- Office 2013,
- All patches (OS & Apps) upto 28th May 2015
We've had a couple of incidents where the one of the RDS boxes started maxing out the CPU, it was determined that each copy of Access running on the server started using as much CPU as it could get, this was confirmed as we moved users off, the remaining users took the remaining CPU, until the last three users were using 75%, 2 users 50% and one users 25% (one instance of Access per core).
To determine if it was a code issue, we just opened a copy of Access with no database, and again the CPU started maxing out a core to 25% for one user.
After a restart, the server returned to normal, no issues with Access, then a couple of days later, the same happened on the other box, again we moved users off, we have tried some checks, looking at eventlogs, and web searches, we've gone through restarting services (print spooler and other user services that seemed likely) with no effect.
AV isn't scanning the ACCDB files.
Although we are currently only running with 100 users, we are planning on moving about another 250 users onto this platform (with an addition of another RDS box).
Although restarting resolves the issue, due to the number of users we are going to be moving onto this solution, we'd like to mitigate this from happening again.
Any assistance would be appreciated, or if you have any questions, please feel free to ask.
Regards,
Chris