Take the 2-minute tour ×
Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

About a year ago, we virtualized most of our small business Windows and Linux infrastructure, which included starting a fresh (not P2V) Server 2008 R2 VM. This performed acceptably until a few weeks ago, when the following changes were made:

  • Xen was upgraded from 5.5 to 5.6FP2
  • The 2k8R2 C: drive was expanded by 10 gigs in XenCenter, and Windows was allowed to expand the drive to include the added space.
  • Pagefiles were relocated from another virtual drive to the expanded C:

Since these changes, we've had a number of occurrences of the server becoming unresponsive. Exhibited behaviors are that the application and website that query the one database this machine hosts would give timeout errors, RDP connections were likely to either never begin, or not render the login screen to authenticate, virtual console control in XenCenter wouldn't be able to authenticate if the GUI had been locked, or if it was open, any interactions (except mouse tracking) would result in errors about the system being unresponsive. My SNMP monitoring still reports the server and SQL service as available, but any attempt to restart the server through proper means (XenCenter, shutdown /i from another machine, the virtual console if I could interact with it) would fail. Only means to solve this is a "force reboot" from XenCenter.

Troubleshooting steps I've taken so far:

  • Increased RAM allocation
  • Moved to alternate Xen host
  • Installed MS KB979149
  • Brought up another virtual drive moved all paging to it
  • Set up a nightly reboot (just yesterday)

Any ideas on what sort of monitoring to begin to answer what is happening, or any known issues that could lead to this?

share|improve this question

2 Answers 2

I'd start by looking at the IO of the LUN that the C drive is sitting on. Sounds like the disk may be overloaded.

share|improve this answer
    
Storage is on a Dell MD3000, and the SAN stores 6 of the 7 VM's in my pool- none of them have IO issues. Wouldn't we expect IO issues on the other hosts too? –  user90307 Aug 29 '11 at 16:38
    
it depends, I'm not sure what sort of throttling Xen would be doing to keep the other VMs up and running. What sort of counters does Zen provide of what the Guest is doing that you can use to see what's spinning out of control on the guest? –  mrdenny Aug 31 '11 at 0:28

Did you upgrade the installation of XenServer tools in the guest? I wonder if some of your problems could be coming from this.

Another question: What method did you use to expand the disk?

I believe that the method stated here is the best: http://support.microsoft.com/kb/325590

If you went to "Computer Management", right clicked on the drive, and chose "Extend Volume". you may have created some I/O overhead for your windows box. I doubt this could cause issues however it is something worth noting in the future.

If it's not outdated XenServer tools in the guest then I guess i have no more ideas :(

share|improve this answer
    
Yes, the Tools were updated to be current. The expansion was done via right-click, Extend volume- I'd never seen the kb325590 method. How do the results differ between that and right-click expansion? –  user90307 Aug 30 '11 at 22:31

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

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