I get this a quite often when I try to deploy compute-only clusters:
2014 Jan 02 09:51:00,489+0000 ERROR SimpleAsyncTaskExecutor-4| com.vmware.bdd.fastclone.impl.FastClusterCloner: Failed to copy object VmCreateSpec [vmName=piv-template] to object VmCreateSpec [vmName=Cluster53-Worker-2]
com.vmware.aurora.exception.VcException: Cannot find the virtual controller for device VirtualLsiLogicController:0:1
Happens both with shared and local storage for the workers. Sometimes a 'Resume deployment' helps, but sometimes I have to do this a couple of times. Anyone any clues? Thanks.
Tom
The error message is misleading I think, as the LsiLogicController is build in the template. It should exist exactly after the VM is cloned. The only possibility is the template is not copied successfully.
Anyway, are you using the default template or a custom CentOS 6 template? Which Version's VC is used, 5.5 or 5.1?
If you're using the custom template, you can check if the template has any problem through manual VM cloning in VC, and attach one disk to the controller: VirtualLsiLogicController:0:1.
I 'm afraid this problem is related with your whole env. If the template is correct, we might need more information from VC and ESXi.
Would you mind to create VC support bundle, and send back to us, together with all Serengeti logs? The VC support bundle as well as host support bundle where the problem happens, will be useful to detect if there is any issues.
You can export system log in VI client using "Expor System Logs".
thanks
Emma
Hi Emma,
To cut some corners short: default templates and VC 5.1. I'm in the process right now of exporting the system logs, but that seems to take a while. While waiting I have been looking at the workers in question and disk (0:1) is always the swap device. The other two disks are on 0:0 and 0:2, so it seems odd that a (virtual) controller related issue would pop up on one disk only.
How do I get the logs to you?
Rgds,
Tom
Tom,
OK, got it. Just remember one thing, did you see any error before this VC exception? Like connection error or socket read time out etc. If the network in your env has problem, it might has some random query issue, but generally that should not be the same problem all the times.
In addition, you can attach files in the reply directly in advanced edit mode.
thanks
Emma
Hi Emma,
I have tried to reproduce the error, but so far all my cluster deployments have gone without this error (have seen other errors, but they were more Pivotal related and I will pots for them separately). I will update this thread if I run into this error again. Thanks.
Tom,
That's OK.
Have fun with Serengeti!
Emma