vSphere Availability

I have a strange strange issue when a couple of my ESXi servers were powered off for a few days, when I booted them back up I couldnt get FT working across a a couple of VM's or I coulnt even use Vmotion to migrate to another host.

I have rebuilt VC and both ESXi hosts, latest versions available. The error i keep getting is a general system error occurred: Failed to initalize migration at source. Error 0xbad00a4. Vmotion failed to start due to lack of CPU or memory resources. Also getting : No Compatible host has sufficnet resources to satisfy the reservation

This error makes no sense at all, there is plenty of spare capacity on both of my ESXi hosts.

Hi, when you say you've rebuilt VC and ther hosts, are there only 2 ESXi servers in this environment? Just trying to get a perspective here. Don't know if it'll help but I did have a weird one with HA on 3.5 (but didn't affect vMotion) recently. It seemed the HA config was corrupt and removing the hosts and recreating the cluster didn't fix it. I had to manually uninstall the two aam agents (the package with vcint in the name has to come off first) and then reconect the hosts which then re-installed the aam packages and directory structure and files which don't get removed when you just disable HA. Oh it's not one of those old ESXi HP dongles is it?

Thanks,

Neil

My problem is back ! Driving me nuts at the moment !.

Yes i only have two ESXi hosts, everything is running the latest versions. First problem was when I had a CPU reservation configured on my VM and wasn't able to migrate. Now I cant migrate anything from within the resource pools even though I have Enable reservation ticked.

If fact now ive deleted the resource pool and still cant even migrate. This has to be a bug surely ? Im trying to migrate to a host which inst evening being utilised using Use Available CPU resources. The host when the VM resides in using hardly any CPU , but memory is 75 % utilised. No dongle is being used.

Ok my problem is when I use Reserve CPU for optimal VMotion performance , it times out .....................

What HA Runtime info show?

---

MCSA, MCTS Hyper-V, VCP 3/4, VMware vExpert '2009

http://blog.vadmin.ru

See attached, but I have now removed all resource groups, I can vmotion but only Perform with available CPU resources.

I cant setup further FT VM's because its not able to migrate the secondary.

Hi,

you know that you don´t have any slots available?

Maybe thats the problem.

Frank

Hi,

look at this paper.

http://www.vmware.com/pdf/vsphere4/r40/vsp_40_availability.pdf

Page 17

Frank

hmm I did see that and wonder, I have 19GHz of CPU and 32Gb of RAM, how can I only have 6 slots with that configuration in my entire cluster?

If its saying I've got 6 slots @ 256. MHz, 1 Virtual CPU, 2177 MB of RAM that's 1.5GHz CPU and 13062 RAM in total, less than half of my total capacity...

This cant be right, I've now powered down every VM expect for 3 !

Here is detailed description what slot is: http://www.yellow-bricks.com/vmware-high-availability-deepdiv/

You need to lower your memory reservation or set slot size manually. 1 GB would be suitable for you I think.

---

MCSA, MCTS Hyper-V, VCP 3/4, VMware vExpert '2009

http://blog.vadmin.ru

im not using any memory reservation what so ever, I cant see how the figures add up ! Only 6 slots between two hosts ?

Double check all VMs. There should be reservation somewhere. Or do you have BIG VMs with multiple CPUs?

---

MCSA, MCTS Hyper-V, VCP 3/4, VMware vExpert '2009

http://blog.vadmin.ru

You can check all your vm´s with only one powershell command.

get-vm

http://communities.vmware.com/community/vmtn/vsphere/automationtools/powercli?ei=9yHgS9yLEZahONTJ2cII&sa=X&oi=spell&resnum=0&ct=result&cd=1&ved=0CB8QBSgA&q=vmware%20powershell%20cli&spell=1

Thats realy amazing.

Frank

ok I had one VM which had a memory reservation, I'll been through every VM now and I have 29 free slots. each VM has 1024 MB and one CPU.

I still cant migrate a VM using "Reserve CPU for optimal perf...................... when I use this option it tells me that " A general system error occurred: Migration failed copying data. Connection closed by remote host, possibly due to timeout."

When I use "Perform with available CPU resources" all is ok...........

When you use the migration with "Reserve CPU for optimal perf, then VMware reservate cpu and memory on the destination host. When there isn´t enouth ressources available, migration fails.

The other option don´t do that. So Migration always succeed.

Frank

Forsure I know that :smileywink: I dont have a resource problem, I have three VM's running!

Look at the attached jpg's , the error doesn't make sense at all.

OK :smileyhappy:

How many Hostfailures have you configured?

Maybe thats the reason why you have only a view slots available

Frank

admission control is disabled, this had been working fine but there is an under lying problem somewhere ?

Hello Virtech,

I believe I can speak about this issue from some bad personal experience. Please follow the instructions below and you will be fine. You don't have to rebuild the hosts not even to reboot them.

1. On the management console (vCenter or vClient) highlight the host on the left side and select the configuration tab on the right side. That opens the host related configuration settings

2. In the right panel (the configuration tab) select "Advanced Settings" The last option

3. In the window that pops up (pic 1) select the "Migrate" option on the left

4. The veri first setting on the right side is Migrate.Enabled the setting's value is 1 by default.

5. Flip the value of that field from "1" to "0" and click OK. You DON'T NEED REBOOT !!!!

6. Wait few seconds and go to the same setting. At this time flip the setting from "0" to "1" and click OK

7. Repeat the steps above on the ssecond host(s) that participate in VMotion

8. Open a beer and enjoy

Sincerely

Ivaylo

I will own my own brewery by the time I get this sorted, tried this and it hasnt made any difference.........................

Hi,

sorry but i think you have to call VMware.

:smileysad:

Frank

there is definitely a bug with ESXi4 , I've just installed ESX4 (full version) and I dont have any issues. Available slots is 72 (thats more like it)

yep, i think youre right.

there are a lot more bugs i think. Login via SSL and Type df -h

But do not wonder why the root partition is nearly full. Thats ESXi

Frank

Have you Open a call yet?

Or are you able to fix it by yourself?

Frank

Any update on this? I'm actually having the exact same issue. It happened after a failed attempt to enable VMware Fault Tolerance. I now have 0 slots available despite having plenty of resources available. I'm using ESX4 not ESXi and am on the latest patches.

I wasnt able to log a support call with VMware and didnt have the time to try and investigate so I rebuilt my environment with ESX4 (full version). Everything is now happy again. There is def a bug with ESXi, and some very strange business when enabling FT....

Sorry I cant be of more help, log a support call if you can with VMware.

As someone mentioned above, for me this turned out to be an Admission Control problem. Once I enabled it and set the amount of host failures then I was able to use VMotion again. VMotion was working for me before as well so I think enabling Fault Tolerance when there weren't enough slots available was what caused my issue.

Hope this helps anyone else who finds this thread.