'bosh cck' creating deployment locks

[work....@gmail.com]

I tried scaling cloud foundry up by adding more runners and deploying it. I have done this before without issue but today the deployment failed and I just could not repeat the deployment again because some of the runners were not responding.

I am doing 'bosh cck' to clear up the unresponsive VMs but 'bosh cck' just fails with the following error:

Error 100: Unable to get deployment lock, maybe a deployment is in progress. Try again later.

I then use 'bosh locks' to check for locks. Wait till the locks are cleared and then try 'bosh cck' again. But 'bosh cck' again creates a deployment lock by itself and errors out with the error above.

Any help or pointers on how to tackle this?

[Etourneau Gwenn]

I thought it was a task lock, I mean avoiding two task to run in the same time.

As cck is a task it can make sens but I am not sure about this.

[fm...@pivotal.io]

The previous task is likely on a state that will eventually timeout but has not done that yet... a common case might be a disconnected host (the unresponsive VMs are on a disconnected host, so bosh can't get to that host to verify the status of the VMs or even detach the disk).

You should first check which tasks are still running (bosh tasks --no-filter )  and then cancel the running tasks (bosh cancel task --task_id) before trying cck.

cheers,

Fred

[work....@gmail.com]

When I start of the 'bosh cck' there are no tasks and no locks. That is where I find the problem is. 'bosh cck' is the one creating locks and it is also erroring out on the lock.

I have done this many times and find this to be happening again and again.

I know it sounds a little weird. It has not happened before.