Host is disabled

olivierlambert

In general, it means XAPI is waiting for something to happen. And in a lot of cases, it's related to mount waiting/PBD.plug of a storage repository.

gawlo

@olivierlambert I wanted to move the vdi from one local repository to another, but it didnt work - now no progress or task is running.

Can i check the xapi status for what he is waiting for? And maybe kill this task?

olivierlambert

Check if all your SR are connected (in SR view in XO, they should be all green)

gawlo

They are all green

olivierlambert

So it might be something else then… Network unreachable somewhere? Hard to tell

gawlo

Could it be a NAS (ISO Repository) who is offline?

olivierlambert

Clearly. But you said everything was green? Anyway, this is blocking the rest.

gawlo

It was yes, but the NAS was offline.
Now XOA shows "server_init (on myserver) 0%" in tasks. Will this exit the maintenance mode?

olivierlambert

What's the result of xe task-list on the host?

gawlo

@olivierlambert Host ist still in maintenance mode...
Result of xe task-list is just blank

gawlo

There is one SR (ISO repos) that is shown offline but it is definitely online. I cannot repair or remove it.

If i try to remove it, xoa starts "Async.PBD.unplug" in the tasks, but after few minutes/hours the task is gone and the SR is still there.

olivierlambert

That's because XAPI tries to unplug something you can't reach. Try with xe pbd-unplug force=true uuid=<UUID>.

gawlo

With this command i get (copy-pasted the uuid)
The uuid you supplied was invalid.
type: PBD
uuid: UUID from SR

Edit: xe task-list = empty

olivierlambert

What UUID are you using? My command was an example.

gawlo

I used this UUID respectively this command "xe pbd-unplug force=true uuid=b7b541e7-a44d-61b7-0db4-55a0f9ce78bb"

olivierlambert

And are you sure this UUID was the PBD UUID? (and not the SR UUID)

gawlo

It was the false UUID....sorry! With the correct UUID it took about 3-4 hours and I ended up with this error:

xe pbd-unplug force=true uuid=44e9e1b7-5a7e-8e95-c5f1-edeebbc6863c
Another operation involving the object is currently in progress
class: SR
object: b7b541e7-a44d-61b7-0db4-55a0f9ce78bb (NAS2 ISO)
<extra>: plug

olivierlambert

I edited your post to use markdown syntax

Anyway, you have something blocking the mount or umount, try to find the process, kill him. Alternatively, lazy umount.

That's what happening when you remove physically a SR before removing it from XAPI

gawlo

Many thanks for the help!!

Unfortunately I still could not solve the problem. Is it advisable to restart the host completely?

EDIT: Found this in the xensource.log:
PBD.destroy locking failed: caught transient failure OTHER_OPERATION_IN_PROGRESS: [ SR; OpaqueRef:d4815da4-f7ca-40b0-ab21-f67411a7abf3; plug ]

I already checked all SRs and PBDs but the UUID can not be found.

Could this be the problem?

olivierlambert

Yes, there's a lock somewhere on the missing SR.