Thread 'Issue in work fetch schedule?'
Message boards : Questions and problems : Issue in work fetch schedule?
Message board moderation
| Author | Message |
|---|---|
|
Send message Joined: 9 Apr 06 Posts: 302 |
18/04/2020 16:18:51 | World Community Grid | update requested by user 18/04/2020 16:18:56 | World Community Grid | Sending scheduler request: Requested by user. 18/04/2020 16:18:56 | World Community Grid | Not requesting tasks: don't need (CPU: job cache full; NVIDIA GPU: ) 18/04/2020 16:18:58 | World Community Grid | Scheduler request completed 18/04/2020 16:20:54 | Einstein@Home | project suspended by user 18/04/2020 16:21:04 | PrimeGrid | project suspended by user 18/04/2020 16:21:09 | SETI@home | update requested by user 18/04/2020 16:21:12 | SETI@home | Sending scheduler request: Requested by user. 18/04/2020 16:21:12 | SETI@home | Not requesting tasks: don't need (CPU: job cache full; NVIDIA GPU: ) 18/04/2020 16:21:14 | SETI@home | Scheduler request completed CPU uses all cores indeed, but GPU sits idle. So, how can be "don't need" if GPU idle ?? ID: 97811 ·  |
|
Send message Joined: 31 Dec 18 Posts: 314 
|
18/04/2020 16:18:51 | World Community Grid | update requested by user Certainly on the WCG side the project does not use GPUs for any of it’s current apps. For SETI, I thought that was in hibernation at the moment? ID: 97814 · |
|
Send message Joined: 5 Oct 06 Posts: 5149
|
Come on, you've been running BOINC as long as I have! Set <sched_op_debug> and let's see what's going on in a bit more detail. You can even do that from the GUI these days - Ctrl+Shift+F ID: 97815 · |
|
Send message Joined: 9 Apr 06 Posts: 302 |
Come on, you've been running BOINC as long as I have! Sorry, Richard, you always was more whole-BOINC oriented than me, so forgive me such repeating ignorance :) OK, to topic. Seems wrong to me: SETI 18/04/2020 19:43:20 | SETI@home | update requested by user 18/04/2020 19:43:22 | SETI@home | sched RPC pending: Requested by user 18/04/2020 19:43:22 | SETI@home | [sched_op] Starting scheduler request 18/04/2020 19:43:22 | SETI@home | Sending scheduler request: Requested by user. 18/04/2020 19:43:22 | SETI@home | Not requesting tasks: don't need (CPU: job cache full; NVIDIA GPU: ) 18/04/2020 19:43:22 | SETI@home | [sched_op] CPU work request: 0.00 seconds; 0.00 devices 18/04/2020 19:43:22 | SETI@home | [sched_op] NVIDIA GPU work request: 0.00 seconds; 0.00 devices 18/04/2020 19:43:24 | SETI@home | Scheduler request completed 18/04/2020 19:43:24 | SETI@home | [sched_op] Server version 709 18/04/2020 19:43:24 | SETI@home | Project requested delay of 1818 seconds 18/04/2020 19:43:24 | SETI@home | [sched_op] Deferring communication for 00:30:18 PrimeGrid 18/04/2020 19:43:37 | PrimeGrid | sched RPC pending: Requested by user 18/04/2020 19:43:37 | PrimeGrid | [sched_op] Starting scheduler request 18/04/2020 19:43:37 | PrimeGrid | Sending scheduler request: Requested by user. 18/04/2020 19:43:37 | PrimeGrid | Requesting new tasks for NVIDIA GPU 18/04/2020 19:43:37 | PrimeGrid | [sched_op] CPU work request: 0.00 seconds; 0.00 devices 18/04/2020 19:43:37 | PrimeGrid | [sched_op] NVIDIA GPU work request: 1.00 seconds; 1.00 devices 18/04/2020 19:43:38 | PrimeGrid | Scheduler request completed: got 0 new tasks WCG 18/04/2020 19:43:29 | World Community Grid | Sending scheduler request: Requested by user. 18/04/2020 19:43:29 | World Community Grid | Not requesting tasks: don't need (CPU: job cache full; NVIDIA GPU: ) 18/04/2020 19:43:29 | World Community Grid | [sched_op] CPU work request: 0.00 seconds; 0.00 devices 18/04/2020 19:43:29 | World Community Grid | [sched_op] NVIDIA GPU work request: 0.00 seconds; 0.00 devices 18/04/2020 19:43:32 | World Community Grid | Scheduler request completed 18/04/2020 19:43:32 | World Community Grid | [sched_op] Server version 701 18/04/2020 19:43:32 | World Community Grid | Project requested delay of 121 seconds What is wrong here: PrimeGrid ASKED work, but SETI and WCG even refuse to ask! Why?... EDIT: And I definitely know that at least SETI is capable to feed this GPU (if tasks are there) so should ask and get nothing in return due to lack of work... EDIT2: more info: E@h constantly asks for NV work on that host and get nothing because of too old GPU (memory requirements missed), PrimaryGrid completed 1 task on this GPU now send zero, but asks for work. ID: 97826 · |
|
Send message Joined: 9 Apr 06 Posts: 302 |
For SETI, I thought that was in hibernation at the moment? It still processing resends so there are some chances to get tasks if ASK, but client refuses to ask - that's the issue. ID: 97827 · |
|
Send message Joined: 5 Oct 06 Posts: 5149
|
That hasn't really helped, has it? We'll probably have to move on to the whole <work_fetch_debug>, but that's horribly verbose. Set it, let it run once, then unset it again. Then, snip that single iteration from start to finish, and post it here. Edit - what version of BOINC are you running, anyway? we put some fairly serious work fetch fixes into v7.16.5 ID: 97829 · |
|
Send message Joined: 9 Apr 06 Posts: 302 |
Here we go BOINC 7.4.42: 19/04/2020 01:59:38 | | [work_fetch] Request work fetch: Core client configuration 19/04/2020 01:59:40 | | [work_fetch] ------- start work fetch state ------- 19/04/2020 01:59:40 | | [work_fetch] target work buffer: 8640.00 + 0.00 sec 19/04/2020 01:59:40 | | [work_fetch] --- project states --- 19/04/2020 01:59:40 | collatz | [work_fetch] REC 0.000 prio 0.000 can't request work: suspended via Manager 19/04/2020 01:59:40 | Milkyway@home | [work_fetch] REC 0.000 prio 0.000 can't request work: suspended via Manager 19/04/2020 01:59:40 | SETI@home Beta Test | [work_fetch] REC 0.000 prio 0.000 can't request work: suspended via Manager 19/04/2020 01:59:40 | SETI@home | [work_fetch] REC 884.959 prio -0.407 can request work 19/04/2020 01:59:40 | World Community Grid | [work_fetch] REC 0.002 prio -1000.015 can request work 19/04/2020 01:59:40 | PrimeGrid | [work_fetch] REC 58.814 prio -1000.034 can request work 19/04/2020 01:59:40 | Einstein@Home | [work_fetch] REC 159.087 prio -1000.172 can request work 19/04/2020 01:59:40 | Rosetta@home | [work_fetch] REC 1071.541 prio -4928.692 can request work 19/04/2020 01:59:40 | | [work_fetch] --- state for CPU --- 19/04/2020 01:59:40 | | [work_fetch] shortfall 0.00 nidle 0.00 saturated 33118.50 busy 21982.47 19/04/2020 01:59:40 | collatz | [work_fetch] share 0.000 blocked by project preferences 19/04/2020 01:59:40 | Milkyway@home | [work_fetch] share 0.000 zero resource share 19/04/2020 01:59:40 | SETI@home Beta Test | [work_fetch] share 0.000 blocked by project preferences 19/04/2020 01:59:40 | SETI@home | [work_fetch] share 1.000 19/04/2020 01:59:40 | World Community Grid | [work_fetch] share 0.000 zero resource share 19/04/2020 01:59:40 | PrimeGrid | [work_fetch] share 0.000 blocked by project preferences 19/04/2020 01:59:40 | Einstein@Home | [work_fetch] share 0.000 zero resource share 19/04/2020 01:59:40 | Rosetta@home | [work_fetch] share 0.000 19/04/2020 01:59:40 | | [work_fetch] --- state for NVIDIA GPU --- 19/04/2020 01:59:40 | | [work_fetch] shortfall 8640.00 nidle 1.00 saturated 0.00 busy 0.00 19/04/2020 01:59:40 | collatz | [work_fetch] share 0.000 19/04/2020 01:59:40 | Milkyway@home | [work_fetch] share 0.000 19/04/2020 01:59:40 | SETI@home Beta Test | [work_fetch] share 0.000 19/04/2020 01:59:40 | SETI@home | [work_fetch] share 0.000 no applications 19/04/2020 01:59:40 | World Community Grid | [work_fetch] share 0.000 no applications 19/04/2020 01:59:40 | PrimeGrid | [work_fetch] share 0.000 project is backed off (resource backoff: 7392.09, inc 19200.00) 19/04/2020 01:59:40 | Einstein@Home | [work_fetch] share 0.000 project is backed off (resource backoff: 1207.61, inc 19200.00) 19/04/2020 01:59:40 | Rosetta@home | [work_fetch] share 0.000 no applications 19/04/2020 01:59:40 | | [work_fetch] ------- end work fetch state ------- 19/04/2020 01:59:40 | | [work_fetch] No project chosen for work fetch 19/04/2020 01:59:46 | SETI@home | update requested by user 19/04/2020 01:59:46 | | [work_fetch] Request work fetch: project updated by user 19/04/2020 01:59:50 | SETI@home | sched RPC pending: Requested by user 19/04/2020 01:59:50 | | [work_fetch] ------- start work fetch state ------- 19/04/2020 01:59:50 | | [work_fetch] target work buffer: 8640.00 + 0.00 sec 19/04/2020 01:59:50 | | [work_fetch] --- project states --- 19/04/2020 01:59:50 | collatz | [work_fetch] REC 0.000 prio 0.000 can't request work: suspended via Manager 19/04/2020 01:59:50 | Milkyway@home | [work_fetch] REC 0.000 prio 0.000 can't request work: suspended via Manager 19/04/2020 01:59:50 | SETI@home Beta Test | [work_fetch] REC 0.000 prio 0.000 can't request work: suspended via Manager 19/04/2020 01:59:50 | SETI@home | [work_fetch] REC 884.959 prio -0.407 can request work 19/04/2020 01:59:50 | World Community Grid | [work_fetch] REC 0.002 prio -1000.015 can request work 19/04/2020 01:59:50 | PrimeGrid | [work_fetch] REC 58.814 prio -1000.034 can request work 19/04/2020 01:59:50 | Einstein@Home | [work_fetch] REC 159.087 prio -1000.172 can request work 19/04/2020 01:59:50 | Rosetta@home | [work_fetch] REC 1071.541 prio -4928.692 can request work 19/04/2020 01:59:50 | | [work_fetch] --- state for CPU --- 19/04/2020 01:59:50 | | [work_fetch] shortfall 0.00 nidle 0.00 saturated 33107.83 busy 21978.53 19/04/2020 01:59:50 | collatz | [work_fetch] share 0.000 blocked by project preferences 19/04/2020 01:59:50 | Milkyway@home | [work_fetch] share 0.000 zero resource share 19/04/2020 01:59:50 | SETI@home Beta Test | [work_fetch] share 0.000 blocked by project preferences 19/04/2020 01:59:50 | SETI@home | [work_fetch] share 1.000 19/04/2020 01:59:50 | World Community Grid | [work_fetch] share 0.000 zero resource share 19/04/2020 01:59:50 | PrimeGrid | [work_fetch] share 0.000 blocked by project preferences 19/04/2020 01:59:50 | Einstein@Home | [work_fetch] share 0.000 zero resource share 19/04/2020 01:59:50 | Rosetta@home | [work_fetch] share 0.000 19/04/2020 01:59:50 | | [work_fetch] --- state for NVIDIA GPU --- 19/04/2020 01:59:50 | | [work_fetch] shortfall 8640.00 nidle 1.00 saturated 0.00 busy 0.00 19/04/2020 01:59:50 | collatz | [work_fetch] share 0.000 19/04/2020 01:59:50 | Milkyway@home | [work_fetch] share 0.000 19/04/2020 01:59:50 | SETI@home Beta Test | [work_fetch] share 0.000 19/04/2020 01:59:50 | SETI@home | [work_fetch] share 0.000 no applications 19/04/2020 01:59:50 | World Community Grid | [work_fetch] share 0.000 no applications 19/04/2020 01:59:50 | PrimeGrid | [work_fetch] share 0.000 project is backed off (resource backoff: 7381.96, inc 19200.00) 19/04/2020 01:59:50 | Einstein@Home | [work_fetch] share 0.000 project is backed off (resource backoff: 1197.48, inc 19200.00) 19/04/2020 01:59:50 | Rosetta@home | [work_fetch] share 0.000 no applications 19/04/2020 01:59:50 | | [work_fetch] ------- end work fetch state ------- 19/04/2020 01:59:50 | SETI@home | [sched_op] Starting scheduler request 19/04/2020 01:59:50 | SETI@home | [work_fetch] request: CPU (0.00 sec, 0.00 inst) NVIDIA GPU (0.00 sec, 0.00 inst) 19/04/2020 01:59:50 | SETI@home | Sending scheduler request: Requested by user. 19/04/2020 01:59:50 | SETI@home | Not requesting tasks: don't need (CPU: job cache full; NVIDIA GPU: ) 19/04/2020 01:59:50 | SETI@home | [sched_op] CPU work request: 0.00 seconds; 0.00 devices 19/04/2020 01:59:50 | SETI@home | [sched_op] NVIDIA GPU work request: 0.00 seconds; 0.00 devices 19/04/2020 01:59:56 | SETI@home | Scheduler request completed 19/04/2020 01:59:56 | SETI@home | [sched_op] Server version 709 19/04/2020 01:59:56 | SETI@home | Project requested delay of 1818 seconds 19/04/2020 01:59:56 | SETI@home | [sched_op] Deferring communication for 00:30:18 19/04/2020 01:59:56 | SETI@home | [sched_op] Reason: requested by project 19/04/2020 01:59:56 | SETI@home | Entry in app_config.xml for app 'setiathome_v8', plan class 'ati_opencl_sah' doesn't match any app versions 19/04/2020 01:59:56 | SETI@home | Entry in app_config.xml for app 'setiathome_v8', plan class 'reserved' doesn't match any app versions 19/04/2020 01:59:56 | | [work_fetch] Request work fetch: RPC complete 19/04/2020 02:00:01 | | [work_fetch] ------- start work fetch state ------- 19/04/2020 02:00:01 | | [work_fetch] target work buffer: 8640.00 + 0.00 sec 19/04/2020 02:00:01 | | [work_fetch] --- project states --- 19/04/2020 02:00:01 | collatz | [work_fetch] REC 0.000 prio 0.000 can't request work: suspended via Manager 19/04/2020 02:00:01 | Milkyway@home | [work_fetch] REC 0.000 prio 0.000 can't request work: suspended via Manager 19/04/2020 02:00:01 | SETI@home Beta Test | [work_fetch] REC 0.000 prio 0.000 can't request work: suspended via Manager 19/04/2020 02:00:01 | SETI@home | [work_fetch] REC 884.959 prio -0.000 can't request work: scheduler RPC backoff (1812.88 sec) 19/04/2020 02:00:01 | Rosetta@home | [work_fetch] REC 1071.541 prio -1.051 can request work 19/04/2020 02:00:01 | World Community Grid | [work_fetch] REC 0.002 prio -1000.015 can request work 19/04/2020 02:00:01 | PrimeGrid | [work_fetch] REC 58.814 prio -1000.052 can request work 19/04/2020 02:00:01 | Einstein@Home | [work_fetch] REC 159.087 prio -1000.222 can request work 19/04/2020 02:00:01 | | [work_fetch] --- state for CPU --- 19/04/2020 02:00:01 | | [work_fetch] shortfall 0.00 nidle 0.00 saturated 33095.78 busy 21974.17 19/04/2020 02:00:01 | collatz | [work_fetch] share 0.000 blocked by project preferences 19/04/2020 02:00:01 | Milkyway@home | [work_fetch] share 0.000 zero resource share 19/04/2020 02:00:01 | SETI@home Beta Test | [work_fetch] share 0.000 blocked by project preferences 19/04/2020 02:00:01 | SETI@home | [work_fetch] share 0.000 19/04/2020 02:00:01 | Rosetta@home | [work_fetch] share 1.000 19/04/2020 02:00:01 | World Community Grid | [work_fetch] share 0.000 zero resource share 19/04/2020 02:00:01 | PrimeGrid | [work_fetch] share 0.000 blocked by project preferences 19/04/2020 02:00:01 | Einstein@Home | [work_fetch] share 0.000 zero resource share 19/04/2020 02:00:01 | | [work_fetch] --- state for NVIDIA GPU --- 19/04/2020 02:00:01 | | [work_fetch] shortfall 8640.00 nidle 1.00 saturated 0.00 busy 0.00 19/04/2020 02:00:01 | collatz | [work_fetch] share 0.000 19/04/2020 02:00:01 | Milkyway@home | [work_fetch] share 0.000 19/04/2020 02:00:01 | SETI@home Beta Test | [work_fetch] share 0.000 19/04/2020 02:00:01 | SETI@home | [work_fetch] share 0.000 no applications 19/04/2020 02:00:01 | Rosetta@home | [work_fetch] share 0.000 no applications 19/04/2020 02:00:01 | World Community Grid | [work_fetch] share 0.000 no applications 19/04/2020 02:00:01 | PrimeGrid | [work_fetch] share 0.000 project is backed off (resource backoff: 7370.54, inc 19200.00) 19/04/2020 02:00:01 | Einstein@Home | [work_fetch] share 0.000 project is backed off (resource backoff: 1186.07, inc 19200.00) 19/04/2020 02:00:01 | | [work_fetch] ------- end work fetch state ------- 19/04/2020 02:00:01 | | [work_fetch] No project chosen for work fetch 19/04/2020 02:00:06 | | Re-reading cc_config.xml 19/04/2020 02:00:06 | | Not using a proxy 19/04/2020 02:00:06 | | Config: simulate 4 CPUs 19/04/2020 02:00:06 | | Config: use all coprocessors ID: 97832 · |
|
Send message Joined: 9 Apr 06 Posts: 302 |
And it seems I understand why SETI asked no work here.... ID: 97833 · |
|
Send message Joined: 9 Apr 06 Posts: 302 |
Thanks, Richard, for narrative questions that lead to understanding where real problem is. In part of BOINC manager - would be good if default log would be little more verbose(or precise) on the reason not giving tasks. To exclude the need in deciphering that reason. In case under question BOINC scheduling is OK. E@h asks for work, doesn't recive due to lack of required device capabilities. SETI doesn't ask cause it's in anonymous platform mode and lack of NV app (it was ATi-based host before GPU burnt). PrimeGrid asks for work and time to time even get it. Seems shortage of GPU tasks there (how it could be for purely mathematics project is beyond my understanding, but it's another question). WCG doesn't ask for work perhaps because there is no NV GPU apps at all (not sure in this point). Thanks again. ID: 97842 · |
|
Send message Joined: 5 Oct 06 Posts: 5149
|
No problem. Glad to be of help. That 'no applications' really leaps out at you, doesn't it? Actually, no, it doesn't. It takes the eye of a programmer to find it amidst all the noise. Although I routinely set <sched_op_debug> on all installations, and given the choice I'd have that active by default, I would adamantly argue against defaulting <work_fetch_debug>. It's huge, it's frequent, it's complex, and it overwhelms everything else. To the average user, it's a waste of space (literally - it bloats their disk storage). It is, however, very valuable in cases like this. I guided you to it, and went to bed. I saw that 'no applications' first thing this morning, but by then you'd sorted it out by yourself. Job done. But while you're here - you said you're running BOINC v7.4.42. That's - erm - ancient. We're in the middle of trying to shake the bugs out of v7.16.5. One new thing in that version - arguably, one of the bugs - is extra massages at startup which might have flagged up this type of problem. People hate them. We still haven't worked out the answer to that one. ID: 97844 · |
|
Send message Joined: 9 Apr 06 Posts: 302 |
No problem. Glad to be of help. That 'no applications' really leaps out at you, doesn't it? Yep :)
Yes, my lead credo now is don't fix what ain't broken. So I refuse to do any updates until will have the need for update. Saw too many cases and once even bricked device because of auto-updates just for the sake of keeping latest version available. That's why such BOINC version. I upgraded to it some time ago as fix to some GPU-related (most probably) bugs in some prev buid and sice then it worked. Actually, it's good sign of quality of that particular release. All those years there was no need to update it! But now SETI lack of work so I can join as beta-tester for BOINC itself. Regarding that message - seems it's part of standard Windows dump. So, not smth coded in BOINC but rather in Windows debug DLL. I would agree with that topic starter, disclose the title of foreground window is quite serious breach in privacy. It's even bigger if one recall that result's stderr will be accessible anonymously, publicly, free from any point of the world. The way I would treat this issue is filter. Filter whole stderr.txt before adding it to client_state or sending back to server. And delete such secuirity breaks proactively. EDIT: hm, seems I answered on wrong link? My answer on this : https://boinc.berkeley.edu/forum_thread.php?id=1357 Will look right one... ID: 97845 · |
|
Send message Joined: 9 Apr 06 Posts: 302 |
Well, I recive such notices too. Adapted to just ignore them. Though they deceive me time to time - I think there is smth really new BOINC would tell me. So, not sure that this "service" as Jord called it is a good one in the first place. And definitely better to have an separate option to disable such notices. And better via GUI. If not, at least to implement same mechanism as here: https://boinc.berkeley.edu/forum_thread.php?id=13574&postid=97534 ID: 97846 · |
Copyright © 2025 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.