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

I am trying to determine the cause of outages i have experienced on an irregular basis with several of my virtual servers which run on vmware esxi4. I have 12 virtual servers spread across 2 esxi host machines and i've only noticed this problem with 3 of them. The virtual servers with problems are one ftp server and two dns servers.

When the outages occur i can log into the vSphere client and use the console of the virtual servers in question, but during the outages (which last anywhere from a minute to near an hour) the virtual servers seem to either experience incredibly slow network transfer speeds in the order of 0-3 kilobits per second or no connection at all. (the datacentre they are in has 100mb connection)

The OS on the machines is debian lenny, and is regularly updated.

I haven't been able to determine the cause of these outages and they really are beginning to bother me so any help in determining the cause of these problems would be much appreciated.

the logfile from esxi contains only

Feb 08 13:16:57.981: mks| SOCKET 8 (105) recv error 104: Connection reset by peer
Feb 08 13:16:57.981: mks| SOCKET 8 (105) destroying VNC backend on socket error: 1
Feb 08 13:16:57.998: mks| MKS: Base polling period is 1000000us
Feb 08 20:28:14.774: mks| MKS: Base polling period is 10000us
Feb 08 20:28:15.194: mks| VNCENCODE 9 encoding mode change: (720x400x16depth,16bpp,1440bytes/line)
Feb 08 20:28:15.228: mks| VNCENCODE 9 encoding mode change: (720x400x16depth,16bpp,1440bytes/line)

repeated several times. Is this simply me connecting and disconnecting to/from the console?

share|improve this question
    
Vmware support anre excellent at log trawling for just this kind of thing, have you put a call in yet? –  Chopper3 Feb 8 '11 at 20:52
    
@Chopper3 unfortunately i have no budget for that so its up to me. I was just hoping maybe someone would be able to point me in the direction of what i should look at. I've been through the logs on both the guests and the hosts and there isn't any fault/error/warning that matches the downtimes. –  m3z Feb 8 '11 at 21:01

1 Answer 1

Here is what I would check into:

  1. Update the ESXi to the latest build (4.1)
  2. Update VMtools on the guests (VMs) to the latest.
  3. Turn off everything that you don't need on the VMs (daemons) - this obviously should be done for any server.
  4. As per VMware's recommendation, validate you have assigned 1vCPU per VM, unless you running applications that specifically able to use multiple CPUs.
  5. Validate that you don't over-commit resources such as memory.
  6. Copy the vmware log for the VM to your PC, usually located in the same directory in the datastore as the VM itself and examine it for strange messages.

I assume you have validated that the hosts are VMware-compatible, otherwise VMware will not support you.

share|improve this answer
    
3. already done - 4. VM's only have 1vCPU - 5. vSphere client reports ~80% mem used, everything else less - 6. done, only things in it for days were as edited into my question above. As for updating, i will look into that tomorrow. Thanks –  m3z Feb 8 '11 at 22:02
    
Turn on higher logging at the properties of the VM. It might add some info. –  Vick Vega Feb 8 '11 at 22:36

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.