Unable to access file since it is locked

I came across this error earlier today and spent the entire day troubleshooting it.

So I wanted to share with you what my resolution was. I was getting this error on a MS cluster setup within my environment. I was working with a Systems Engineer; we were converting an existing MS cluster from using vmdks for the quorum and msdtc shared drives to RDMs. Once Node A was configured, I attached the drives and powered up Node B. Node B would make it to 95% and then provide the error (shown above).

I checked and double-checked my settings. Second SCSI controller set to LSI Logic Parrallel Controller Type with Physical SCSI Bus Sharing selected (VMs were on separate hosts). SCSI IDs set to SCSI 1:1, 1:2, and 1:3 for the new RDM Drives. The RDMs set to virtual compatibility mode.

After numerous attempts of re-configuring as I was doubting that I did it right. I consulted Google – which lead me to an older white paper (one for ESX 3.5). This wasn’t too bad, as the steps are pretty much the same for our ESX 4.0 Update 2 environment. Since the white paper verified that I was doing the right thing, I dug deeper into Google for an answer. I came across this vmware KB Article (#10051) that provides an indepth detail of the error. So now I’m worried. I start troubleshooting according to the KB article, and it did allow me to find something rather odd. Even though the VM was physically sitting in one Datastore, the ESX host believed it to be in another datastore. A quick svmotion corrected this, but it was REALLY odd.

Almost on the verge of giving up, I consult with a co-worker. He starts going down the list of tasks to do that is eerily familiar – they were the same steps from the white paper. Once they were triple-checked, he asks me the one question that I’ve asked the SE three or four times previously. Has the cluster been completely unconfigured. Turns out the cluster was completely removed on Node A, but Node B had not been evicted. Once the cluster cleanup was done on Node B, the two nodes booted up fine with the RDMs attached without an error. Eight hours later, the cluster is reconfigured and running fine.

The moral of the story. Rule #1 is always right. Do not trust what the user tells you. 🙂

Enable EVC when vCenter is running as a VM

I wanted to enable EVC in my lab environment, but found that I couldn’t because my VC was a VM within that same cluster. VC refused to enable it because there was still at least one VM powered on within the cluster. Duh.

Of course, I broke out my handy-dandy do-it-all tool (yep, google) and proceeded to locate a workaround. Lo and behold, I found a useful link. Thanks go out to Bert at http://virtwo.blogspot.com.

He had instructions for how to do this within an ESX 3.5U2 environment. My environment is a complete ESX 4.1 environment consisting of two ESX 4.1 Hosts. Bert’s instructions work fairly well in my environment.

Here’s what I did for my environment:

1) Evacuate one ESX host in your cluster and place it in maintenance mode.
2) Move it out of the cluster, right under your datacenter object.
3) Exit the host out of maintenance mode.
4) Manually migrate (VMotion via drag-and-drop for example) your VirtualCenter VM (running on an ESX host in the cluster) to the ESX host that is now outside the cluster. Do this with all other VMs that were still running in the cluster.
5) Enable EVC on your cluster.
6) Migrate the VMs back to the ESX host within the cluster (again, drag and drop).
7) Place the ESX host back into the cluster.

Now, I have a happy cluster running EVC.

Here’s the link for Bert’s blog entry.
There’s also a KB Article (#1013111) on this subject as well. You can find it here.