Message boards :
CMS Application :
no new WUs available
Message board moderation
Previous · 1 . . . 9 · 10 · 11 · 12 · 13 · 14 · 15 . . . 20 · Next
| Author | Message |
|---|---|
 Send message Joined: 29 Aug 05 Posts: 1006 Credit: 6,272,224 RAC: 353  |
It looks like I was a bit optimistic last night... However, our WMAgent is now fully operational again, and tasks/jobs are available, Thanks for your patience during the outage.  |
|
Send message Joined: 2 May 07 Posts: 2101 Credit: 159,817,517 RAC: 132,770  
|
no CMS tasks in the download queue Is back again. |
|
Send message Joined: 27 Sep 08 Posts: 807 Credit: 652,444,676 RAC: 280,103 |
No issue for me, I see 166 and I have ca 800 jobs in my queue. |
|
Send message Joined: 29 Aug 05 Posts: 1006 Credit: 6,272,224 RAC: 353 |
No issue for me, I see 166 and I have ca 800 jobs in my queue. Yes, the post puzzled me too. BTW, as far as terminology goes, you have tasks in your queue. As I tried to explain elsewhere tonight, a CMS@Home BOINC task is an instantiation of our VM which then attempts to run CMSSW jobs from the condor pool. |
|
Send message Joined: 2 May 07 Posts: 2101 Credit: 159,817,517 RAC: 132,770
|
no new WUs available |
|
Send message Joined: 29 Aug 05 Posts: 1006 Credit: 6,272,224 RAC: 353 |
I've just noticed that the number of running jobs seems to be falling. No obvious reasons why from my preliminary checks. I'll be in my office in an hour or so, where I can do more detailed investigations. |
|
Send message Joined: 2 May 07 Posts: 2101 Credit: 159,817,517 RAC: 132,770
|
atm the flow of new tasks is ok. There are Volunteers with problems in running a CMS-Task (for example short run less than 10 sec.) This mean a lot of tasks from CMS have no restart for a normal using. Ivan, we hoping you have a wingman for your input of new tasks in the near future. Thank you for your work! |
|
Send message Joined: 29 Aug 05 Posts: 1006 Credit: 6,272,224 RAC: 353 |
Yeah, things are looking OK at the moment, but the delay in some monitoring reports means I can't be totally sure that we are still on track. I'll keep an eye on it. I do have a wingperson, but at the moment she is not actively involved in workflow injection. One good thing that's emerged in the last month is that I misremembered my contract details, and I should still be here until March 2023, health permitting (I'm now booster jabbed, and had the 'flu shot as well, so hopefully I'll get through this winter OK 🤞). |
|
Send message Joined: 2 May 07 Posts: 2101 Credit: 159,817,517 RAC: 132,770
|
All the best for you, Ivan :-)). |
|
Send message Joined: 29 Aug 05 Posts: 1006 Credit: 6,272,224 RAC: 353 |
OK, the glitch seems to have been temporary, but it may have affected post-production jobs that run on the T3_CH_CMSAtHome cluster at CERN. I'll check with Laurence tomorrow if the situation there still seems iffy. |
 Send message Joined: 15 Jun 08 Posts: 2413 Credit: 226,454,837 RAC: 132,412  |
Computers are getting tasks but the subtask queue seems to be empty since 11:00 UTC this morning. |
|
Send message Joined: 14 Jan 10 Posts: 1280 Credit: 8,493,530 RAC: 2,126   |
Computers are getting tasks but the subtask queue seems to be empty since 11:00 UTC this morning. The mechanism to stop creating BOINC workunits, when no sub-tasks are avaiable for the VMs, seems to work: Unsent 0 |
|
Send message Joined: 29 Aug 05 Posts: 1006 Credit: 6,272,224 RAC: 353 |
Computers are getting tasks but the subtask queue seems to be empty since 11:00 UTC this morning. Oops, sorry 'bout that! I must have forgotten to check the status of the queues this morning... I had some connectivity problems today, had to reset my WiFi modem a few times; guess I forgot which sites I hadn't been able to visit. New workflow injected, jobs should become available again "Real Soon Now" (© Jerry Pournelle). |
|
Send message Joined: 15 Jun 08 Posts: 2413 Credit: 226,454,837 RAC: 132,412 |
The CMS subtask queue seems to be empty since 8:24 UTC. May be a result of another major outage last night. |
|
Send message Joined: 29 Aug 05 Posts: 1006 Credit: 6,272,224 RAC: 353 |
The CMS subtask queue seems to be empty since 8:24 UTC. Yes, there is a problem with the WMAgent and no new jobs are being created. I've e-mailed the people who are able to restart it. There does seem to have been a network problem last night, similar to what we had last Friday night. There was a large spike in the number of jobs failing with error 8002, which usually means they haven't been able to contact any of the frontier servers (i.e. conditions database). |
|
Send message Joined: 29 Aug 05 Posts: 1006 Credit: 6,272,224 RAC: 353 |
The agent has been restarted and jobs are available again. I'm told there was a network problem at CERN last night but, correlating different sources, I think our agent problem actually started several hours before that. |
|
Send message Joined: 29 Aug 05 Posts: 1006 Credit: 6,272,224 RAC: 353 |
I'm having problems submitting a new batch of jobs tonight, and I can't make much sense of the error message. I estimate that we have about 12 hours before we start to run out of jobs, so I'll try again in the morning. |
|
Send message Joined: 29 Aug 05 Posts: 1006 Credit: 6,272,224 RAC: 353 |
OK, THAT problem was eventually solved. Now we are having problems going forward. A change has unexpectedly meant that volunteer machines are not getting production jobs, but instead post-production jobs that are meant to go to a dedicated VM cluster at CERN. We suspect some sort of typo/misconfiguration, but it's hard to track down the right people at this time of year. Setting No New Tasks or switching to other projects is, as always, recommended in these cases. |
|
Send message Joined: 29 Aug 05 Posts: 1006 Credit: 6,272,224 RAC: 353 |
We have found the source of the problem, and jobs are available again. |
|
Send message Joined: 18 Dec 15 Posts: 1688 Credit: 103,858,099 RAC: 121,542
|
Happy New Year to everybody :-) Unfortunately, the new year does not begin very well for CMS: NO NEW TASKS :-( |
©2024 CERN