Message boards : Number crunching : Keep getting no tasks available errors
Message board moderation

To post messages, you must log in.

Previous · 1 · 2

AuthorMessage
Profile Magic Quantum Mechanic
Avatar

Send message
Joined: 24 Oct 04
Posts: 1114
Credit: 49,501,728
RAC: 4,157
Message 42449 - Posted: 13 May 2020, 8:07:37 UTC
Last modified: 13 May 2020, 8:08:43 UTC

I am on trying to set all of my Cern work up since my high-speed just started an new month after 1am as usual

WHY do I get this all the time?
5/13/2020 12:59:29 AM | LHC@home | No tasks are available for Theory Simulation
5/13/2020 12:59:29 AM | LHC@home | Tasks for AMD/ATI GPU are available, but your preferences are set to not accept them
5/13/2020 12:59:29 AM | LHC@home | Project requested delay of 6 seconds


I keep getting that | Project requested delay of 6 seconds and some at 61 seconds over and over......and of course the server is not out of these tasks.
And it doesn't repeat the request over after that 6 seconds so I do it manually over and over until I just give up...(like right now)....this is something new added and it needs to be removed from Updates requested by User.
ID: 42449 · Report as offensive     Reply Quote
Crystal Pellet
Volunteer moderator
Volunteer tester

Send message
Joined: 14 Jan 10
Posts: 1268
Credit: 8,421,616
RAC: 2,139
Message 42451 - Posted: 13 May 2020, 9:39:48 UTC - in response to Message 42449.  

6 seconds delay are from LHC@home and the 61 seconds delay are from LHC-dev.

As I understand the work-queuing: There is only a small buffer in memory with queued tasks to distribute.
This memory buffer could be filled with only ATLAS/Sixtrack tasks, so when requesting Theory there is nothing to distribute at the moment.
This will repeat until enough tasks are distributed and some "Theory's ready to send" are placed in that memory buffer.

Solution: Make 5 memory buffers for each application one.
ID: 42451 · Report as offensive     Reply Quote
computezrmle
Volunteer moderator
Volunteer developer
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 15 Jun 08
Posts: 2386
Credit: 222,892,787
RAC: 138,196
Message 42452 - Posted: 13 May 2020, 9:55:30 UTC - in response to Message 42449.  
Last modified: 13 May 2020, 9:57:59 UTC

Nothing new (most likely).

There are 2 possible reasons:
1. You accidentally checked "Run native if available?".
2. The project runs 2 main servers concurrently and specific tasks are out on one of them for unknown reason.

In case of (2.):
Since your client decides which project server it sends the request to is made DNS based and the DNS record for lhcathome.cern.ch is valid for 60 s it makes not much sense to press the update button within those 60 s even if the server tells you a shorter delay.
(Well, except if the server meanwhile has refilled it's shared memory queue, but that's a time expensive procedure)
There's also no guarantee that you get the other server after a DNS refresh since the IPs are always in random order.

<edit>
I agree that CP mentioned another possible reason.
</edit>
ID: 42452 · Report as offensive     Reply Quote
Crystal Pellet
Volunteer moderator
Volunteer tester

Send message
Joined: 14 Jan 10
Posts: 1268
Credit: 8,421,616
RAC: 2,139
Message 42454 - Posted: 13 May 2020, 11:01:56 UTC - in response to Message 42452.  
Last modified: 13 May 2020, 11:24:21 UTC

Within a few hours I saw Sixtrack tasks 'in progress' grew from 18k to 23k, so a lot of Sixtrack's are created and placed in the distribution queue, although Unsent is 0 (zero) all the time.
Also to mention that much more tasks are distributed, cause many Sixtrack's are short-runners.

That's probably causing 'No tasks available' for other applications every now and then.
ID: 42454 · Report as offensive     Reply Quote
Profile Magic Quantum Mechanic
Avatar

Send message
Joined: 24 Oct 04
Posts: 1114
Credit: 49,501,728
RAC: 4,157
Message 42463 - Posted: 13 May 2020, 21:06:01 UTC

As I said......I manually tell it to send work AFTER those 6 seconds.....NOT before.
This is only about the new version of Theory here.
AND I never have been set for "Native" since I am 100% Windows so why would I ever have it set that way?
Why you guys seem to act like I don't know what I am doing or am new to this always amazes me.

I just aborted the Theory tasks and switched to only Sixtracks
I would add more but one of you would use your mod power for taking care of that.....but it sure would be nice if you actually had been running these for over 9 years (VB) every day at home with your own pc's instead of running one or two over at -dev and then pretending like you know how it works and then try to tell me I am doing something wrong.....which I have to draw with crayons with a reply every time.
(and I remember everything that ever happened at T4T and neither one was at the Alpha-Atlas site either)
ID: 42463 · Report as offensive     Reply Quote
Jim1348

Send message
Joined: 15 Nov 14
Posts: 602
Credit: 24,371,321
RAC: 0
Message 42465 - Posted: 13 May 2020, 21:44:29 UTC - in response to Message 42463.  

I have given up trying to deal with it logically, or by random intervention.
It is only giving me two native ATLAS. That leaves more cores for Rosetta.
It is divine intercession.
ID: 42465 · Report as offensive     Reply Quote
Harri Liljeroos
Avatar

Send message
Joined: 28 Sep 04
Posts: 674
Credit: 43,148,944
RAC: 16,010
Message 42953 - Posted: 3 Jul 2020, 10:24:12 UTC

I am facing this problem today. All subprojects state that there are tasks available on server status page, but I cannot get any. The response is just 'No tasks available'.
ID: 42953 · Report as offensive     Reply Quote
tullio

Send message
Joined: 19 Feb 08
Posts: 708
Credit: 4,336,250
RAC: 0
Message 42957 - Posted: 3 Jul 2020, 16:48:29 UTC

I am getting both Atlas tasks and Theory tasks , no CMS tasks.
Tullio
ID: 42957 · Report as offensive     Reply Quote
Harri Liljeroos
Avatar

Send message
Joined: 28 Sep 04
Posts: 674
Credit: 43,148,944
RAC: 16,010
Message 42958 - Posted: 3 Jul 2020, 17:51:47 UTC - in response to Message 42957.  

My other Host has now received a full cache of tasks but the other one hasn't received any since I made my post here. I have now selected all subprojects and accept tasks also from other applications if none are available for preferred applications. Let's see if that is the solution for this.
ID: 42958 · Report as offensive     Reply Quote
Jim1348

Send message
Joined: 15 Nov 14
Posts: 602
Credit: 24,371,321
RAC: 0
Message 42959 - Posted: 3 Jul 2020, 21:37:09 UTC - in response to Message 42958.  

I have never found a certain solution for that. At the moment, I am getting both native ATLAS and CMS with no problems.
Sometimes the server likes you, sometimes it doesn't.
ID: 42959 · Report as offensive     Reply Quote
Harri Liljeroos
Avatar

Send message
Joined: 28 Sep 04
Posts: 674
Credit: 43,148,944
RAC: 16,010
Message 42961 - Posted: 3 Jul 2020, 23:17:54 UTC - in response to Message 42959.  

I managed to get one sixtrack task on my main cruncher and that's it. I downloaded a few Einstein CPU tasks just to keep the system warm (I usually do Einstein only on GPUs), there are two of those running now and they are finished in an hour or so.

My other Host has kept it's cache full during this evening, so the server just doesn't like my other computer ;-P. OK, off to bed now and I'll see in the morning if anything has changed.
ID: 42961 · Report as offensive     Reply Quote
Harri Liljeroos
Avatar

Send message
Joined: 28 Sep 04
Posts: 674
Credit: 43,148,944
RAC: 16,010
Message 42969 - Posted: 5 Jul 2020, 10:50:13 UTC

Yesterday the problem moved to the other host :-(
But today both are getting new tasks again. So fingers crossed that this status stands.
ID: 42969 · Report as offensive     Reply Quote
neocortix_0

Send message
Joined: 23 May 20
Posts: 1
Credit: 1,580,407
RAC: 0
Message 42971 - Posted: 6 Jul 2020, 20:15:55 UTC

I have a bunch of aarch64 hosts that have been starved for tasks for several days. Should I expect the situation to change, or are you all done supporting computation on ARM64?
ID: 42971 · Report as offensive     Reply Quote
Crystal Pellet
Volunteer moderator
Volunteer tester

Send message
Joined: 14 Jan 10
Posts: 1268
Credit: 8,421,616
RAC: 2,139
Message 42974 - Posted: 7 Jul 2020, 5:44:39 UTC - in response to Message 42971.  

I have a bunch of aarch64 hosts that have been starved for tasks for several days. Should I expect the situation to change, or are you all done supporting computation on ARM64?
LHC ran out of available jobs - SixTrack
See the server status page: https://lhcathome.cern.ch/lhcathome/server_status.php
ID: 42974 · Report as offensive     Reply Quote
Previous · 1 · 2

Message boards : Number crunching : Keep getting no tasks available errors


©2024 CERN