Postponed: VM job unmanagable, restarting later.

Message boards : Questions and problems : Postponed: VM job unmanagable, restarting later.
Message board moderation

To post messages, you must log in.

AuthorMessage
Profile adrianxw
Avatar

Send message
Joined: 2 Oct 05
Posts: 400
Denmark
Message 99916 - Posted: 13 Jul 2020, 6:40:32 UTC

>>> Postponed: VM job unmanagable, restarting later.

I've seen this before, and have a work unit in this state right now, it has been there since yesterday afternoon. What is the logic of this state, when does it restart?
Wave upon wave of demented avengers march cheerfully out of obscurity into the dream.
ID: 99916 · Report as offensive
Les Bayliss
Help desk expert

Send message
Joined: 25 Nov 05
Posts: 1654
Australia
Message 99918 - Posted: 13 Jul 2020, 7:25:21 UTC - in response to Message 99916.  

Doing a web search on that error message brings up quite a few results, so it appears to be fairly common.

This one from LHC may help: VM Job Unmanageable
ID: 99918 · Report as offensive
Profile adrianxw
Avatar

Send message
Joined: 2 Oct 05
Posts: 400
Denmark
Message 99926 - Posted: 13 Jul 2020, 12:30:45 UTC
Last modified: 13 Jul 2020, 13:14:11 UTC

Not really. The other jobs from the project are working on an identically configured machine.

2438692 1351883 52 12 Jul 2020, 12:02:39 UTC 13 Jul 2020, 4:55:17 UTC Completed, waiting for validation 59,171.72 58,934.53 pending NWChem long v0.11 (vbox64_t1)
windows_x86_64

And other examples running without issue now. There is a lot of memory in these machines, I have allowed 4GB per core, ie. both machines, (4GHz i7's), are kitted out with 16GB. The tasks are running okay on this machine, which is the one I use most. The task in that state, (it still is incidently), is only running against other BOINC projects. The two systems have similar portfolios of projects, only significant difference is the failing machine has a dodgy graphics card so I have suspended GPU work on it.

<edit>
Elsewhere, I have been advised to dump VirtualBox 6 and go back to 5.2.38. The job that was stuck started again after 24 hours and is running now, they have ENORMOUS time allocations, (sometime in October), so if it sticks for 24 hours every now and again, I guess no harm is done.
Wave upon wave of demented avengers march cheerfully out of obscurity into the dream.
ID: 99926 · Report as offensive
Profile adrianxw
Avatar

Send message
Joined: 2 Oct 05
Posts: 400
Denmark
Message 99967 - Posted: 15 Jul 2020, 7:14:35 UTC - in response to Message 99926.  

When a job is suspended, it is stuck for 24 hours. Is it possible to change that time? The reason I want to do this is because I want to replace virtual box with the older version, which I had the installer for on disk from earlier. Since it is not obvious which projects or tasks are using virtual box, I set no new jobs for all projects, with the thought I'd deinstall VB when they were finished and install the older version, to try to get around the suspension issue. I currently have a single task on both machines here which are suspending regularly.
Wave upon wave of demented avengers march cheerfully out of obscurity into the dream.
ID: 99967 · Report as offensive
Profile Dave
Help desk expert

Send message
Joined: 28 Jun 10
Posts: 2518
United Kingdom
Message 99969 - Posted: 15 Jul 2020, 9:09:37 UTC

4GB/core is the minimum I would accept on a new machine for running BOINC. CPDN testing branch has had tasks that require a peak of 5GB/core and LHC and some others have some tasks that require over 4GB/task. It may be worth checking what the peak memory usage is of these tasks.
ID: 99969 · Report as offensive
Profile adrianxw
Avatar

Send message
Joined: 2 Oct 05
Posts: 400
Denmark
Message 99976 - Posted: 15 Jul 2020, 10:39:02 UTC

When monitoring my system usage, I rarely see more than 60% memory in use at any one time.
Wave upon wave of demented avengers march cheerfully out of obscurity into the dream.
ID: 99976 · Report as offensive
Profile adrianxw
Avatar

Send message
Joined: 2 Oct 05
Posts: 400
Denmark
Message 99979 - Posted: 15 Jul 2020, 13:11:26 UTC

Downgraded Virtual Box to 5.2.38 to see if that actually helps.
Wave upon wave of demented avengers march cheerfully out of obscurity into the dream.
ID: 99979 · Report as offensive
Profile Dave
Help desk expert

Send message
Joined: 28 Jun 10
Posts: 2518
United Kingdom
Message 99980 - Posted: 15 Jul 2020, 14:52:01 UTC - in response to Message 99976.  

When monitoring my system usage, I rarely see more than 60% memory in use at any one time.


With the CPDN tasks that used a fraction over 5GB each, running three at once on my laptop (4cores, 8GB RAM) it was only very occasionally I needed swap as they tended not to peak at the same time. Am I right in thinking the VM has protected RAM allocated to it? I have only dabbled occasionally with VB so am not sure. Using one or two fewer cores and increasing the memory allocated to the VM may be worth trying but I am sure there are people better qualified than myself to answer that.
ID: 99980 · Report as offensive

Message boards : Questions and problems : Postponed: VM job unmanagable, restarting later.

Copyright © 2024 University of California.
Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.