admin
Immortal
Immortal

DRS not working! Arghhhhhhh

Got 5 hosts in a cluster, and stress testing them, so I have put 8x W2k3 VM's on one host each with 16GB RAM and 4vCPUs and created a rule so the must stay together and the DRS setting is set to Manual for VMotion so they cannot DRS off, on this host I have put another VM --no rules so should be able to DRS anywhere, and is setup as default cluster settings, all VM's can VMotion manually.

When I CPU stress the 8 big VMs the Host CPU goes to 100% as expected and stay on the host as expected, but the lone VM does not DRS off to one of the other hosts on the cluster even though the host is on it knees.

Any ideas, am I missing something as like I say all servers are VMotion ready and working, all networking is working, in fact everything is working except DRS across all the hosts --I am completely lost. Smiley Sad

DRS is enabled by the way and all hosts are fuly licenced. Weirdly though if I have another VM on the said host powered down and power it up the VM automatically powers up on another host, as if DRS is working.

This happens on vCenter 3.5 Update 3 with ESX Hosts 3.5 Update 3 & vCenter 3.5 Update 4 with ESX Hosts 3.5 Update 4

THIS IS DRIVUNG ME NUTS AS EVERYOTHER TIME I HAVE DONE THIS THE LONE VM MIGRATES BECAUSE OF DRS, NOTHING HAS CHANGED.

Tags (2)
0 Kudos
12 Replies
oreeh
Immortal
Immortal

DRS rules are a bit "buggy".

Try multiple rules with only 2 VMs per rule.

And make sure DRS is set to aggressive.

0 Kudos
Troy_Clavell
Immortal
Immortal

maybe i'm not understanding your question completely, but if you have a DRS cluster set to manual, no vMotions will be issued, only recommended. Are there any DRS recommendations? Also, what does the "percent of entitled resources delivered" chart look like? If it's all they way over to the left, it may be hard to generate a drs event other than putting your host int maintenance mode.

0 Kudos
admin
Immortal
Immortal

The DRS cluster is set to fully automated and aggressive, but you also have the option to drill down, and create rules to pin individual vms in to groups such as keep together, and another option with regards to DRS behavior such as default/manual etc of which I have done. I have just removed all rules/behavior and still no joy.

0 Kudos
oreeh
Immortal
Immortal

Looks like you aren't stressing the host enough.

Do the VMs migrate to another host when you put a host in maintenance mode? If yes, DRS is working.

0 Kudos
admin
Immortal
Immortal

all 16 cores of the host are 100% the server is getting well hammered. did the same test 6 weeks ago and it worked as expected.

0 Kudos
admin
Immortal
Immortal

What about HA? Is it possible that HA constraints are preventing it?

0 Kudos
weinstein5
Immortal
Immortal

You say the VMs are vmotion - ready have you tried to manually vmotion?

If you find this or any other answer useful please consider awarding points by marking the answer correct or helpful

If you find this or any other answer useful please consider awarding points by marking the answer correct or helpful
0 Kudos
admin
Immortal
Immortal

VMotion manually works, with regards to HA, I have checked the box and that is it + I have checked "enable VM's to power on if the violote etc..."

0 Kudos
nja
Contributor
Contributor

I think i had the same problem some time ago. Maybe i can find anything in my documentation.

Did you check all the Ressource Pools configuration? Some entrys in the logfile?

0 Kudos
admin
Immortal
Immortal

Solved it

0 Kudos
vpalan
Contributor
Contributor

Hi,

how you resolved? could you please share with us.. Smiley Happy

0 Kudos
vpalan
Contributor
Contributor

Hi, Am planning to implement DRS & HA in our environment. Could you share known issue while implementing DRS & HA?

i am using 3.5 and vc 2.5..

thanks in advances.

0 Kudos