Message boards :
Number crunching :
No WUs right now?
Message board moderation
Previous · 1 · 2 · 3 · 4 · 5 · Next
Author | Message |
---|---|
Send message Joined: 28 Feb 15 Posts: 253 Credit: 200,562,581 RAC: 0 |
In situation where server have generated more than about 2kk WU's feeder sometimes stuck for a while with feeding computers. I think I have a clue. Universe apparently asks for new work every time you upload results, even though new work may not be needed for your buffer size. That may overload the server. There seems to be a somewhat similar problem on MilkyWay, though the symptoms may be slightly different. But there is a discussion on it with some suggestions as to the cure. https://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=4424 |
Send message Joined: 30 Jun 16 Posts: 41 Credit: 309,815,029 RAC: 0 |
Well I've moved 2 PC's on to other projects after waking this morning to find them sitting idle. WU's definitely aren't being pushed out. |
Send message Joined: 28 Feb 15 Posts: 253 Credit: 200,562,581 RAC: 0 |
I think I have a clue. Universe apparently asks for new work every time you upload results, even though new work may not be needed for your buffer size. That may overload the server. By the way, I see the same behavior on WCG. It also asks for new work whenever you upload results. But I don't see it on LHC, TN-Grid or Rosetta. They don't ask for new work if the cache is full. For example, TN-Grid Platform Not requesting tasks: don't need (CPU: not highest priority project; NVIDIA GPU:) So it is a server configuration difference? |
Send message Joined: 26 Sep 16 Posts: 3 Credit: 264,690,611 RAC: 0 |
There's a problem getting new work: ------------ business as usual ---- 30.07.2019 02:31:27 | Universe@Home | Sending scheduler request: To fetch work. 30.07.2019 02:31:27 | Universe@Home | Reporting 8 completed tasks 30.07.2019 02:31:27 | Universe@Home | Requesting new tasks for CPU 30.07.2019 02:31:29 | Universe@Home | Scheduler request completed: got 0 new tasks 30.07.2019 02:31:29 | Universe@Home | Project has no tasks available ------------ another PC: ------------ business as usual ---- 01.08.2019 07:08:00 | Universe@Home | Sending scheduler request: To fetch work. 01.08.2019 07:08:00 | Universe@Home | Requesting new tasks for CPU 01.08.2019 07:08:01 | Universe@Home | Scheduler request completed: got 0 new tasks 01.08.2019 07:08:01 | Universe@Home | Project has no tasks available -------------- message provoked by request a few min ago: ---------- 01.08.2019 09:05:21 | Universe@Home | update requested by user 01.08.2019 09:05:21 | Universe@Home | Sending scheduler request: Requested by user. 01.08.2019 09:05:21 | Universe@Home | Requesting new tasks for CPU 01.08.2019 09:05:24 | Universe@Home | Scheduler request completed: got 0 new tasks 01.08.2019 09:05:24 | Universe@Home | Project has no tasks available --------- All the best! |
Send message Joined: 6 Feb 19 Posts: 41 Credit: 131,325,319 RAC: 0 |
I am outta tasks for Universe... Server says tasks available..... but no tasks will download.... I even reset / detached / re-attached..... No tasks downloaded.... |
Send message Joined: 6 Feb 19 Posts: 41 Credit: 131,325,319 RAC: 0 |
Well there are WU's that are downloading. Just not to any of my machines.... 3352778 WU's @ 10:20am EST 3348807 WU's @ 10:55am EST 3346856 WU's @ 11:35am EST 3344527 WU's @ 12:20pm EST So total available WU's is consistently decreasing in number.... But I can't get any.... |
Send message Joined: 6 Feb 19 Posts: 41 Credit: 131,325,319 RAC: 0 |
Well there are WU's that are downloading. Just not to any of my machines.... 3352778 WU's @ 10:20am EST 3348807 WU's @ 10:55am EST 3346856 WU's @ 11:35am EST 3344527 WU's @ 12:20pm EST So total available WU's is consistently decreasing in number.... But I can't get any.... |
Send message Joined: 24 May 19 Posts: 16 Credit: 53,469,167 RAC: 0 |
I think I have a clue. Universe apparently asks for new work every time you upload results, even though new work may not be needed for your buffer size.Whether it asks for new work when submitting results depends on whether it needs more in order to fulfil your client configuration. Although I am mostly getting: Tue 30 Jul 2019 14:32:43 BST | Universe@Home | Requesting new tasks for CPUat the moment, I did once get this: Sun 28 Jul 2019 19:27:44 BST | Universe@Home | Sending scheduler request: Requested by user.because I run several other projects too and they had sufficient work pending in existing jobs. But most of the time I'm getting no Universe@HOME jobs. It was OK until the recent server upgrade. |
Send message Joined: 28 Feb 15 Posts: 253 Credit: 200,562,581 RAC: 0 |
Whether it asks for new work when submitting results depends on whether it needs more in order to fulfil your client configuration. That is the way it is supposed to work, but has been asking for work after every report of completed tasks until a few minutes ago. Now (as of 01:48 UTC) I am beginning to get: Sending scheduler request: To report completed tasks. That is the first time I have seen it for this project. It looks like it is working OK for the moment. We will see if it stays working. Good luck. |
Send message Joined: 18 May 17 Posts: 16 Credit: 3,215,924 RAC: 0 |
Yeah, I haven't gotten WUs for a few days now... |
Send message Joined: 21 Jun 15 Posts: 1 Credit: 3,151,586 RAC: 0 |
Project reset worked for me. |
Send message Joined: 6 Feb 19 Posts: 41 Credit: 131,325,319 RAC: 0 |
You need to have proper versions of the libraries installed in your system: CXXABI_1.3.9 libstdc++.so.6: version `GLIBCXX_3.4.21' I searched for this but I couldn't find it. I have 3 Ubuntu machines that have been updated but still no WU's..... |
Send message Joined: 28 Feb 15 Posts: 253 Credit: 200,562,581 RAC: 0 |
I have 3 Ubuntu machines that have been updated but still no WU's..... It is not random. For some reason, if it likes your machine, you will get work for a few weeks, and then not. So there is something going on. If they can ever figure out what it is, everyone should get work all the time. |
Send message Joined: 4 Feb 15 Posts: 846 Credit: 144,180,465 RAC: 0 |
Yesterday I just switch on 3 computers for the project. Two Linux based x86 and one ARM machine. Slowly, all the computers get work. Krzysztof 'krzyszp' Piszczek Member of Radioactive@Home team My Patreon profile Universe@Home on YT |
Send message Joined: 24 May 19 Posts: 16 Credit: 53,469,167 RAC: 0 |
You need to have proper versions of the libraries installed in your system: CXXABI_1.3.9 libstdc++.so.6: version `GLIBCXX_3.4.21'The absence of such a library won't prevent you getting jobs downloaded - it will just prevent them running onve the have been. So that wouldn't be related to why several of us are seeing a marked reluctance for the server to serve out jobs when asked. |
Send message Joined: 6 Feb 19 Posts: 41 Credit: 131,325,319 RAC: 0 |
The three machines that I am NOT receiving WU's for have: Ubuntu 18.04.2 LTS [4.9.185-44|libc 2.27 (Ubuntu GLIBC 2.27-3ubuntu1)] Ubuntu 18.04.2 LTS [4.9.185-44|libc 2.27 (Ubuntu GLIBC 2.27-3ubuntu1)] Ubuntu 18.04.2 LTS [4.4.143|libc 2.27 (Ubuntu GLIBC 2.27-3ubuntu1)] The other machines that are receiving WU's are: Ubuntu 18.04.2 LTS [4.14.120-160|libc 2.27 (Ubuntu GLIBC 2.27-3ubuntu1)] 4.4.143 4.14.5-92 It does seem that it's something random..... I almost want to say version issues, but I have one machine with 4.4.14 that receives WU's and one with the same version that doesn't receive.... Those two are the same manufacturer different models. |
Send message Joined: 28 Feb 15 Posts: 253 Credit: 200,562,581 RAC: 0 |
Try changing the work buffer size. When I was using the default size of 0.1 + 0.5 days, it would sometimes run out of work. But increasing that to 0.5 + 0.5 days has helped, and I have not run out yet. It does not actually cure the problem, I still see the "no work available" message, but it covers it up well enough for now. |
Send message Joined: 18 Apr 19 Posts: 1 Credit: 1,142,000 RAC: 0 |
Like lots of us, I have been plagued by the message saying no work is available since the system moved to new servers. The problem is there, and not with us. I found a temporary workaround by disconnecting from the project and resetting it. Then when I reconnected, I got a batch of work units. Now, I'm out again. I run five, now four projects since removing universe. I don't have the time to spend doing nursemaid to the system. The four remaining projects just tick along with no attention from me, as I accumulate about 90,000 points per day. I won't be back unless someone drops me a note, and says it is working properly again. I'm not going to jump through hoops to donate my time. My system is a Corsair One I14o running an i7 9700K at about 4.4 G. I have 32 Gig memory, and an RTX 2080. My email is Ray.Clare@yahoo.com. |
Send message Joined: 6 Feb 19 Posts: 41 Credit: 131,325,319 RAC: 0 |
I have had my buffers set at +5 +1 indefinitely on all machines all using the same global_prefs_override.xml file. I have 3 machines that can't get WU's. I am going to try to reset the project next and see if that helps. |
Send message Joined: 30 Jun 16 Posts: 41 Credit: 309,815,029 RAC: 0 |
My PC's that have a buffer of 0 all run out of work. I can only guess it has to do with how often work is being fetched? |