Hi All,
Sorry for the strange title but its not so easy to explain whats happening.
Basically all vm's on my esxi server suffer from the same problem. When I work in a VM the following happens:
Sometimes I can work great for a few minutes
Then the VMWare freezes (maybe swapping problem) when you look in taskmanager you see that all processes are not responding
Then all of a sudden its gone again
and loop
I'm not a big expert on esxi so its very possible I've made a huge misconfiguration. I could of course try to explain how I've configured my esxi server but perhaps this behavior sounds familiar to someone and perhaps the discussion can lead to possible areas of investigation?
I hope someone has an idea and thanks for your answer.
How much RAM does your ESXi server have and how much is being used by your virtual machines?
While I was running vSphere 4.0 I had 8Gb of RAM - I would try to leave 2Gb for vSphere and the rest for virtual machines (even though according to my summary memory usage I still had some free).
Welcome to the Forums - I agree with mikelane that it sounds like you are having memory contention issues and your VMs are utilizing their per vm vmkernel swap files - one thing ave you installed VMware Tools in all your VMs?
If you find this or any other answer useful please consider awarding points by marking the answer correct or helpful
Hi guys thanks for the welcome and your help.
The general setup of the server is this:
5 Resource pools
Each resource ppol has 1 VM
Each VM has 2 GB memory availble
Now all 5 vm's are running and when I look at the summary tab of the server I see I have a Huge amount of reservers that are NOT used.
The numbers:
CPU Usage 291 MHZ (Capacity 12*1,995 GHZ)
Memory Usage 2991, 00 MB (Capacity 20475,56 MB)
So I have memory enough I'd think... should I increase the VM's memory?
For the record All vm's have vmware tools installed
If I need to provide more info or screendumps let me know.
Thanks,
Joost
Personally I have never used resource pools on a per VM basis, especially when the resource envelope of the pool so closely matches what is defined for just one VM. IMO there is no point in setting a pool up this way.
Do the VMs run fine without the resource pools as you have them currently set?
http://www.vmware.com/pdf/vsphere4/r40/vsp_40_resource_mgmt.pdf Page 35.
Hi Mike,
This setup is just another test to see if it helps.
And no when I have the vm's directly under the server root the problem is the same.
Other setups I've tried (including your suggestion are)
1. No resource pools at all
2. All vm's in one resource pool
3. Half of the vm's in one the other vm's in a second resource pool...
Further info is that 3 vm's I've created from scratch on the esxi machine the other too where vm's I've imported with workstation to esxi server.
And all 5 suffer from the problem with freezing vm's... (my original thought was that perhaps only the imported machines would have problems... but its all of them)
Any more ideas?
Ps I'll read that pdf document you gave me now too.
Thanks
I assume that your datastores are local and not on the network?
I'm also assuming that this behaviour exhibits itself when all VMs are idle?
Do you get the issue with just one VM running?
My instinct would be to start looking at the hardware - test memory with memtest and check the temp of the CPU in the BIOS etc. Then perhaps test the PSU even.
It doesn't make sense that you are having issues with the resources that you have available to your VMs.
Could also be HDD related. What kind of storage/HDDs are you using? Do you use a RAID controller? With or without battery backed write cache?
Even though I think that's not the issue, take a look at http://kb.vmware.com/kb/1018206 (Poor virtual machine application performance on HP hardware)
André
Hi Mike,
Thanks for your answers... my answers are inline.
I assume that your datastores are local and not on the network?
Yes this is true I have 4 1,5 TB disks but for now all 5 vm's are on 1 disk? Would it be an idea to put them on different disks?
I'm also assuming that this behaviour exhibits itself when all VMs are idle?
I'm not sure how I would check this if the vm's are idle? But as an example I have big programs I need to compile and those kind of programs give good progress information so its easy to spot when the vm freezes.. then I usually pull up task manager and then I see all other programs are not responding too.
Do you get the issue with just one VM running?
I just tested this. I've shut all but one VM and then I started the compiler to build my solutions.. there is no change on one machine it freezes too.
My instinct would be to start looking at the hardware - test memory with memtest and check the temp of the CPU in the BIOS etc. Then perhaps test the PSU even.
Ok this is a more difficult area for me but I'll look with our local hardware guy at this question tomorrow and report back the results.
It doesn't make sense that you are having issues with the resources that you have available to your VMs.
hear, hear... I also thought that this machine would be more then capable of running these few vm's...
Hi André,
Thanks for your answer.
To be honest I didn't buy nor setup this ESXI server and the hardware so I have to check this tomorrow to ask what the answer to your question is.
For now I've included an attachment with the esxi properties for all 4 HDD. For the record the vm's are on the Datastore disk. (maybe this gives you some info)
Hi Guys,
I did one more additional test.
I've moved 2 vm's to there own separate datastore/HDD. (amazing detail... copying took 1 hours per 30GB (disk size) machine....)
Anyhow after the move I've started the machines and did the tests again... And again the system freezes.
Keep you posted on the hardware checks I'll try todo today.
Thanks
Hi André,
I've checked your advice in the bios and I did have to change the power options on this server.
However after the reboot the problem still persists...
About HDD to my surprise this server doesn't have a raid controller so I basicaly have 4 SATA disks... But with a machine with these specs could it be the problem?
I really hope you guys have more ideas about the freezing of VM's problem.
Thanks for all the help!