Message boards :
Number crunching :
Panic Mode On (70) Server problems?
Message board moderation
Previous · 1 . . . 5 · 6 · 7 · 8 · 9 · Next
Author | Message |
---|---|
kittyman Send message Joined: 9 Jul 00 Posts: 51468 Credit: 1,018,363,574 RAC: 1,004 |
Well, I see nothing really got fixed today with dear ol' Synergy in it's new duties. The crickets are still slouching, even though kibble bowls are not full. I have sloooooooooly gained a bit on cache, but still am about 3000 WUs shy of the limits, and have several rigs doing GPU only, because of the silly server bug that seems to avoid sending any CPU work unless the GPU is up to the limits. And so, the CPUs sit idle. Vexing bit of a perfect storm. But the kitties pad along. "Freedom is just Chaos, with better lighting." Alan Dean Foster |
cliff Send message Joined: 16 Dec 07 Posts: 625 Credit: 3,590,440 RAC: 0 |
Well I think I found a way around the problem of misscommunication between boinc and its peer. Hit the update button until it succeeds. I had 19 tasks to report, didnt want more WU and the update failed.So I hit update again [ignoring the timeout] and it failed again, so I hit update again and it made contact ok, reported ok. It looks to my admittedly untrained eye as if there is a problem with the router at berkeley, perhaps memory or a software glitch.Or of course the pipe feeding the site. Anyway my tuppence worth. Cheers Cliff, Been there, Done that, Still no damm T shirt! |
kittyman Send message Joined: 9 Jul 00 Posts: 51468 Credit: 1,018,363,574 RAC: 1,004 |
Well I think I found a way around the problem of misscommunication between boinc and its peer. I think that Synergy is just overloaded, or the Apache settings are not quite right yet.....dunno, just a guess. Synergy is trying to do a yeoman's duty here.... "Freedom is just Chaos, with better lighting." Alan Dean Foster |
cliff Send message Joined: 16 Dec 07 Posts: 625 Credit: 3,590,440 RAC: 0 |
Hi Mark, Well I had another 2 results to report and it went into its usual routine, so I did what I posted I'd done last time and once again 3rd time lucky. If its synergy being overloaded then I guess this wont be sorted out until bane is reconfigured:-/ And methinks that may take Matt a fair bit of time since he's also going to be dealing with the new servers too. Regards, Cliff, Been there, Done that, Still no damm T shirt! |
kittyman Send message Joined: 9 Jul 00 Posts: 51468 Credit: 1,018,363,574 RAC: 1,004 |
Hi Mark, Hopefully the new download server shall take some load off of Synergy.....or however they choose to distribute the work load. Certainly I would think that the new server would handle more than just downloads, just as Synergy is now handling much more than it was originally commissioned for. I suspect good things are indeed on the horizon....and not just the headlight of an approaching train. LOL. "Freedom is just Chaos, with better lighting." Alan Dean Foster |
cliff Send message Joined: 16 Dec 07 Posts: 625 Credit: 3,590,440 RAC: 0 |
Dunno so much about those good things, it seems we have a massive solar storm enroute to us that is due to hit later on this evening. Apparently the northern lights will be seen at lower latitudes than usual and power grids may be affected as will some flight paths. Other than the potential power outage I guess when the new kit comes online things will be better, once its fully configured of course.. So about 2 weeks? or so... Regards, Cliff, Been there, Done that, Still no damm T shirt! |
kittyman Send message Joined: 9 Jul 00 Posts: 51468 Credit: 1,018,363,574 RAC: 1,004 |
Dunno so much about those good things, it seems we have a massive solar storm enroute to us that is due to hit later on this evening. I just posted an article about the solar flares in the Cafe. Hope it is not as disruptive as they are saying it could be. "Freedom is just Chaos, with better lighting." Alan Dean Foster |
Wiggo Send message Joined: 24 Jan 00 Posts: 34747 Credit: 261,360,520 RAC: 489 |
I'm sorry but I still can't complain. Cheers. |
shizaru Send message Joined: 14 Jun 04 Posts: 1130 Credit: 1,967,904 RAC: 0 |
Well, I see nothing really got fixed today with dear ol' Synergy in it's new duties. The crickets are still slouching, even though kibble bowls are not full. Thank you for that. From an old post I made, I can see I've had this problem for at least 12 days now... Suprised I missed it being mentioned. It was driving me crazy 'cause I was sure I messed something up when I signed up for Beta. |
LadyL Send message Joined: 14 Sep 11 Posts: 1679 Credit: 5,230,097 RAC: 0 |
Well, I see nothing really got fixed today with dear ol' Synergy in it's new duties. The crickets are still slouching, even though kibble bowls are not full. You could try using the website setting to get no GPU work for some time, so your CPU cache fills and then reenable GPU workfetch. I'm not the Pope. I don't speak Ex Cathedra! |
David S Send message Joined: 4 Oct 99 Posts: 18352 Credit: 27,761,924 RAC: 12 |
I think that Synergy is just overloaded, or the Apache settings are not quite right yet.....dunno, just a guess. Synergy is trying to do a yeoman's duty here.... As stated on the SSP, the purpose of the feeder is "Fills up the scheduler work queue with workunits ready to be sent. The scheduler is usually too busy handling client transactions to maintain such a queue itself." So is that effect being negated by the scheduling server, scheduler process, and feeder all being on the same machine? Whatever, my i7 has gained some on its quota. It's now over 700 out of its limit of 800, up from 6xx yesterday and around 500 last week. It's also devoid of Einsteins (probably paying off debt to Seti accrued when work wasn't available). David Sitting on my butt while others boldly go, Waiting for a message from a small furry creature from Alpha Centauri. |
kittyman Send message Joined: 9 Jul 00 Posts: 51468 Credit: 1,018,363,574 RAC: 1,004 |
Well, I see nothing really got fixed today with dear ol' Synergy in it's new duties. The crickets are still slouching, even though kibble bowls are not full. Thank you so much for the suggestion, worked like a charm! The kitties now have their CPUs crunching away again. Now....you might persuade D.A. to amend the scheduler programming so this work around is not required. Instead of totally ignoring the slower resource on a computer, perhaps it could try to give just a few percent to the slower resource while still heavily favoring the faster GPUs. Instead of starving the slower CPUs altogether. Just a thought...LOL. (Now I just have to remember to set it back before I go to work...or I'll really have to kick myself...) "Freedom is just Chaos, with better lighting." Alan Dean Foster |
cliff Send message Joined: 16 Dec 07 Posts: 625 Credit: 3,590,440 RAC: 0 |
Dont tell me the sun has hidden his head again? Anyway feel free to whinge as much as you please, nearly everyone else is:-) How long do you think it will take to start putting out AP's again? Days,Weeks, Months. Thats as good as any place to start. Cheers, Cliff, Been there, Done that, Still no damm T shirt! |
Kevin Olley Send message Joined: 3 Aug 99 Posts: 906 Credit: 261,085,289 RAC: 572 |
Now that sounds like a very good idea, if the download server cache is getting clogged with VLAR's and the server is only trying to hand out GPU WU's that could slow things down, we need the servers running as smoothly as possible. Kevin |
kittyman Send message Joined: 9 Jul 00 Posts: 51468 Credit: 1,018,363,574 RAC: 1,004 |
Of course. Don't get me wrong, the idea of prioritizing workfetch for the best computer resource is a good one. It just should not be done with the TOTAL exclusion of repeated requests for CPU work. "Freedom is just Chaos, with better lighting." Alan Dean Foster |
rob smith Send message Joined: 7 Mar 03 Posts: 22202 Credit: 416,307,556 RAC: 380 |
As far as I understand it the cruncher decides what it is going to ask for and the server decides if it can satisfy that request. Thus if you are requesting "GPU only", and there are no WU suitable you will get the dreaded "no tasks available" message, or one of its variants. That said, if you are asking for "any WU I can crunch" preference will be given to satisfying that particular crunchers "best" processor's needs. At the server level there is no such thing as a "GPU WU", but there are types of WU that a particular GPU cannot handle, this is managed by a combination of the cruncher and the server. In general crunchers running "more modern" versions of BOINC will be sending back their current cache states so the servers know if the limits are being infringed or not. Assuming that a cruncher is running a reasonably recent version of BOINC the servers will only be reacting to requests for WU that can be crunched, and only enough to satisfy either the imposed limits, or the locally defined cache, and only WU that the cruncher is capable of processing will be delivered... Bob Smith Member of Seti PIPPS (Pluto is a Planet Protest Society) Somewhere in the (un)known Universe? |
kittyman Send message Joined: 9 Jul 00 Posts: 51468 Credit: 1,018,363,574 RAC: 1,004 |
As far as I understand it the cruncher decides what it is going to ask for and the server decides if it can satisfy that request. Not quite true.... My top host was making work requests for CPU AND GPU. And yet getting NOTHING for the CPU at all. The CPUs were idle. It was not until I unchecked GPUs in the preferences that Boinc made work requests for CPU only, and the work then started coming in. The mixed requests were not getting any work for the CPU whatsoever. "Freedom is just Chaos, with better lighting." Alan Dean Foster |
Kevin Olley Send message Joined: 3 Aug 99 Posts: 906 Credit: 261,085,289 RAC: 572 |
+1 Kevin |
Khangollo Send message Joined: 1 Aug 00 Posts: 245 Credit: 36,410,524 RAC: 0 |
My top host was making work requests for CPU AND GPU. And yet getting NOTHING for the CPU at all. The CPUs were idle. It was not until I unchecked GPUs in the preferences that Boinc made work requests for CPU only, and the work then started coming in. The mixed requests were not getting any work for the CPU whatsoever. I'm observing the same annoying behavior on my host. Combined CPU+GPU requests only start getting some CPU work when cache is near full, otherwise I'm getting GPU work only. When astropulse was still operational, I sometimes got 1-2 AP CPU workunits during those combined requests, even when cache was nearly empty. Also, to make matters much much worse, if you let your host to "bounce" off the limits, cache never gets full enough and you get no CPU work at all. |
musicplayer Send message Joined: 17 May 10 Posts: 2430 Credit: 926,046 RAC: 0 |
I have noticed that with the two most recent versions of BOINC Manager I am using, 6.10.58 and particularly 6.12.34, that when attaching to the server (be it Seti@home or PrimeGrid), the server gives me both CUDA-based as well as non-CUDA tasks (PrimeGrid has more options than Seti@home for setting preferences for the types of tasks being run than Seti@home). Anyway, I get both types and especially 6.12.34 towards PrimeGrid gives me a whole bunch of new tasks. It is when I first set the CUDA-option active after first finishing setting up things, making sure it works, these tasks eventually starts running. My guess is that the server looks at the version of BOINC Manager being used first, then next what the capabilities of the PC being used are, including the processor. |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.