DRS No Longer Making Recommendations

[sreboulet]

mRecently I've noticed that my DRS is no longer making recomendations. I migrated several VM's off of one of my ESX servers to reboot it (we have a cluster of 3), and noticed when I checked on it several hours later that despite the ESX being up and responding, it still had no VM's on it, nor were there any recommendations to migrate any VM's despite one of the ESX servers showing alarms due the volume of VM's running on it.

I unchecked and rechecked the DRS/HA settings in the cluster, and allowed another 1/2 hour, and it still had no recomendations.

One thing I did notice was if I started a VM that was shutdown and hosted by the overloaded ESX, it would start up on the ESX that was 'empty'. So this seems to be DRS/HA IS working, but only for VM's at startup, and not doing anything for VM's that are already running.

I'm sure this wasn't the case before. I've rebooted the ESX-control server several times, but that doesn't seem to help.

It's currently set as: Fully Automated, "Apply recommendations with three or more stars"

Anyone have any ideas where else to look?

Thanks!

-stephen-

[sbeaver]

If the other hosts in the cluster are not getting any real kind of workout then migrations will not happen. You set the automation on 3 stars raise it to 5 starts if you want to try to get it to move.

Steve Beaver

VMware Communities User Moderator

*Virtualization is a journey, not a project.*

[sreboulet]

Thanks! I forgot I went through my VM's and lowered the memory requirements based on the logs. I guess I lowered the requirements enough that it was happy as it was with the settings at 3 stars. Once I changed the star settings, it started making recomendations.

-stephen-

[XFoS]

I am having a similar issue to the OP.

DRS used to work fine - load was spread nicely between the two hosts in the cluster based on memory requirements (CPU is lower than RAM), with each host sitting at around 35%. VI3.5u1 was in use.

We upgraded to VI3.5u2 (VC first, then VI Clients, then ESX hosts). To upgrade the first host, we put it into maintenance mode, and all VMs moved off it (as expected). We took the ESX host down, uppgraded it (with the ESX 3.5u2 CD) and brought it back up, took it back out of maintenance mode, and waited for DRS to re-balance the load, but nothing happened. Figuring perhaps this was due to a mis-match of ESX versions, we placed the other host in maintenance mode, all VMs moved off it to the other host, downed the host, upgraded it, brought it back up, took it out of maintenance mode, and an hour and a half later, I'm still waiting for recommendations or moves.

The cluster is set to move automatically, any recommendation of 2 stars or higher. It's worked well in the past, but I can't even get it to make recommendations at all today unless I place a host in maintenance mode.

Any ideas gratefully accepted.