Message boards :
Number crunching :
Panic Mode On (23) Server problems
Message board moderation
Previous · 1 . . . 7 · 8 · 9 · 10 · 11 · 12 · 13 · Next
Author | Message |
---|---|
Michele Di Masi (KE6GRN) Send message Joined: 26 Aug 09 Posts: 3 Credit: 40,244 RAC: 0 |
Hi All, I have not gotten any assigments for almost two days. Is there a problem with SETI? Has anybody else been having problems getting assigments? Are the SETI servers down? If so does anybody know when they will come back on? Thanks, Michele Di Masi |
Patrick Pearson Send message Joined: 19 May 99 Posts: 6 Credit: 12,206,167 RAC: 15 |
Part of the problem of lack of work may be the number of work units that users are hogging on their systems. I have just checked back on a WU one of my computers finished in July and found that the person who is also crunching that unit has over 1200 work units waiting to be processed. There is no way he is going to get them done in time. How do users manage to download and hog so many units and so leave little or no work for many of us? Is there not a ceiling on the number of units any one computer can download at any one time? Surely it should be linked in some way to the processing capacity of that particular machine or it just wastes everyone's time. |
1mp0£173 Send message Joined: 3 Apr 99 Posts: 8423 Credit: 356,897 RAC: 0 |
Can someone please explain why the technical folks have not given any indication of what is currently wrong? It's been this way for quite a while, now. Certainly long enough for a comment or explanation. I understand that they are overworked, BUT...it is just commomn courtesy. Matt is the one who generally writes, so if we aren't getting updates, it probably means he's on vacation. Eric has posted in a couple of threads, but as you point out, they're short on staff, and if Matt is on vacation then they're down by 1/3rd. Besides the project is doing what they've always promised. They've always promised that there will be times when there isn't enough to go around. |
David @ TPS Send message Joined: 30 Sep 04 Posts: 70 Credit: 11,323,275 RAC: 0 |
Patrick: Not being sure where you got that data, I have probably that many "pending" on my user page but they are spread over probably 15 boxes and 50 cores. My cache is set for 10 days. I could suspend SETI for a week or 2 and still (proably) finish them in time, but I wont. I am just increasing the shares of my backup projects. Boinc will throw SETI into High Priority if it feels it needs to. If you look at individual computers, I have probably 50 or less on each box. Dave |
HAL9000 Send message Joined: 11 Sep 99 Posts: 6534 Credit: 196,805,888 RAC: 57 |
Patrick: I have seen several users making post of "i just downloaded 800, 1500, or 1200 tasks" after an outage and the servers comming back up. Some of these are users with just 1 or 2 machines. SETI@home classic workunits: 93,865 CPU time: 863,447 hours Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[ |
Salvador Lopez Send message Joined: 20 May 99 Posts: 12 Credit: 2,183,995 RAC: 0 |
I'm exactly in the same position since yesterday I receive 02/09/2009 09:53:41 a.m. SETI@home Message from server: (Project has no jobs available) |
David @ TPS Send message Joined: 30 Sep 04 Posts: 70 Credit: 11,323,275 RAC: 0 |
WOW! Guess my machines aren't that greedy! |
arkayn Send message Joined: 14 May 99 Posts: 4438 Credit: 55,006,323 RAC: 0 |
This tells you what we are dealing with currently. Results ready to send 1 5 0m Current result creation rate 0.1940/sec 0.5151/sec 5m Currently there is no work being created so there is nothing to be sent out. Keep an eye on this link if you see those messages. http://setiathome.berkeley.edu/sah_status.html |
Gundolf Jahn Send message Joined: 19 Sep 00 Posts: 3184 Credit: 446,358 RAC: 0 |
...There is no way he is going to get them done in time. How do users manage to download and hog so many units and so leave little or no work for many of us? Why not? On some CUDA GPUs, a WU takes about 5 minutes to finish (some more some less). That makes 288 per day per GPU. Some hosts have 4 GPUs; that makes 1152 tasks per day. So, the host in question probably has a cache of one or two days (for bigger caches, BOINC manager seems to get instable). Gruß, Gundolf Computer sind nicht alles im Leben. (Kleiner Scherz) SETI@home classic workunits 3,758 SETI@home classic CPU time 66,520 hours |
Alinator Send message Joined: 19 Apr 05 Posts: 4178 Credit: 4,647,982 RAC: 0 |
Can someone please explain why the technical folks have not given any indication of what is currently wrong? It's been this way for quite a while, now. Certainly long enough for a comment or explanation. I understand that they are overworked, BUT...it is just commomn courtesy. LOL... Especially when they are doing things like bringing a new Master BOINC Database Server online, rolling out ATi GPU support, getting started on a new school year, and probably a dozen other items I'm not thinking of right offhand! ;-) The part I'm wondering about is what's the story with Jocelyn being disabled? This means that all queries have to go through Mork, which isn't going to help its performance if past history is any indication. I suppose it could be they are trying to see how much grunt Mork has. The alternative is something we probably really don't want to think about (went in-Sidious on us)! :-D Alinator |
Gundolf Jahn Send message Joined: 19 Sep 00 Posts: 3184 Credit: 446,358 RAC: 0 |
...(went in-Sidious on us)! :-D rotflmao [edit]When again was Sidious taken out of service?[/edit] |
Alinator Send message Joined: 19 Apr 05 Posts: 4178 Credit: 4,647,982 RAC: 0 |
A week or ten days ago, IIRC. Alinator |
Alinator Send message Joined: 19 Apr 05 Posts: 4178 Credit: 4,647,982 RAC: 0 |
Or more simply: 1.) Yes 2.) Yes 3.) No, but currently as useful as pockets in underwear. 4.) No, but we could always start a pool on when the jam will clear! :-D Alinator |
1mp0£173 Send message Joined: 3 Apr 99 Posts: 8423 Credit: 356,897 RAC: 0 |
This tells you what we are dealing with currently. If the creation rate is not zero, then work is being created. The numbers are pretty small, so there is likely more demand than there is work, and many requests will go unfilled. |
1mp0£173 Send message Joined: 3 Apr 99 Posts: 8423 Credit: 356,897 RAC: 0 |
Do we know that SETI is working on ATI support? We know that BOINC is working on that, but first BOINC has to know how to detect and schedule another GPU type, and then someone has to develop the app. I don't know how Mork performs, and I'm not going to second-guess the reason for the change. Someone will probably tell us eventually. |
Alinator Send message Joined: 19 Apr 05 Posts: 4178 Credit: 4,647,982 RAC: 0 |
I don't know about application support for MB and AP, but we do know we are being used for testing the BOINC framework support for ATi GPUs, which was what I referring to in this case. So as far as that goes, I don't remember a case where something new got added to BOINC which didn't break three other things or cause all kinds of unexpected problems. ;-) In any event, since I run multiple projects the net impact on my hosts overall is zero regardless of what happens here! :-D Alinator |
HAL9000 Send message Joined: 11 Sep 99 Posts: 6534 Credit: 196,805,888 RAC: 57 |
Also there is the chance of ATI & Nvidia cards being mixed in one machine. I'm sure someone will have that configuration. I would for testing at least. BOINC then has to track x # of cpus, y # of nvida gpus, & z # of ati gpus. Might get messy in there. SETI@home classic workunits: 93,865 CPU time: 863,447 hours Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[ |
52 Aces Send message Joined: 7 Jan 02 Posts: 497 Credit: 14,261,068 RAC: 67 |
2 newbie cents about being a Wingman Hostage: My machine (a NetBook w/Atom and 3G -- yes, mobile all day long) is lucky to crank out 1 WU a day, but despite that, it seems I'm finding myself lately waiting up to 6+ days for a Wingman (Anvilman) with some heaven sent dream machine to pipe in with their Concur so that my tiny Acct can be credited. And I agree, I too see Wingman queues with hundreds of In-Progress items, and WORSE, they're not always being tackled in DATE order (and sometimes even have in-progress for deadline expired items). I'm glad that what takes me 30,000 seconds only takes them 120, but I've no accurate idea of when that 120 might begin. I know the queuing system is complex and evolved to optimize the complex needs of the project and the conflicting needs of participants. And having a 30+ year predictive scope makes SETI the wrong project for people gratified by real-time measurable progress, especially as a quorum is apparently required to keep the TI honest in their search for ETI, but perhaps the rules can be changed up a bit so those massive contributing heavy lifters (and I fully accept their contrbution far outstrips mine thus they're entitled to a bigger slice of the pie -- even though others call this hogging or land grabbing) don't hold hostage us multitude of little guys (the larger community of public support which is equally beneficial to the project from an awareness & political perspective). #1. Perhaps try and pair wing partners a bit closer on Avg turn around time. Don't pair a 1-dayer with 6-dayer. Keep it within 100%. #2. Or perhaps change to CREDIT the work performed assumptive that the Concur will come thru (which I bet it does 99.998% of the time), and debit it later if that assumption proves false (ya know, just like a bank balance shows checks that have not yet cleared). You can govern this to historical daily WU's for that individual. #3. Or perhaps slightly skew the points awarded to FAVOR the faster turn-around guy (still dependant on across the board results Concur). Maybe swipe 10% of the delaying partner's eventual credit for the WU for each day of wait time since initial result submission (gets a bit tricky if quorums of 3 or 4 are required -- but what a great Game Theory exercise). ... or maybe it stays the way it is, and that's just part of the price of participation. Cheers ! |
jrusling Send message Joined: 8 Sep 02 Posts: 37 Credit: 4,764,889 RAC: 0 |
Looks like the splitters are back online and splitting. It is probably going to be a while before we can get caught back up though. http://boincstats.com/signature/-1/user/18390/sig.png |
HAL9000 Send message Joined: 11 Sep 99 Posts: 6534 Credit: 196,805,888 RAC: 57 |
2 newbie cents about being a Wingman Hostage: My machine (a NetBook w/Atom and 3G -- yes, mobile all day long) is lucky to crank out 1 WU a day, but despite that, it seems I'm finding myself lately waiting up to 6+ days for a Wingman (Anvilman) with some heaven sent dream machine to pipe in with their Concur so that my tiny Acct can be credited. And I agree, I too see Wingman queues with hundreds of In-Progress items, and WORSE, they're not always being tackled in DATE order (and sometimes even have in-progress for deadline expired items). I'm glad that what takes me 30,000 seconds only takes them 120, but I've no accurate idea of when that 120 might begin. I've always thought recent average credit was stupid myself. If I am trying to judge performance of two machines. I will really only know if those two machines are preforming completed tasks. I think an option you would want would be something along the lines of "only send me tasks that need validation". Which I imagine would be rather complex on the server end & doesn't really benifit the project in any way. SETI@home classic workunits: 93,865 CPU time: 863,447 hours Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[ |
©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.