Problems with Update


log in

Advanced search

Message boards : Number crunching : Problems with Update

Author Message
jravin
Send message
Joined: 25 Mar 02
Posts: 929
Credit: 97,061,312
RAC: 86,632
United States
Message 1301226 - Posted: 2 Nov 2012, 11:24:00 UTC
Last modified: 2 Nov 2012, 11:55:28 UTC

For more than a day my 2 machines keep getting timeouts when trying to do an Update, either automatically from BOINC or manually. I get:

11/2/2012 7:39:23 AM | SETI@home | Scheduler request failed: Timeout was reached
11/2/2012 7:39:26 AM | | Project communication failed: attempting access to reference site
11/2/2012 7:39:29 AM | | Internet access OK - project servers may be temporarily down.

every time, yet uploads work OK. I have changed nothing in this span, yet both machines have nearly 400 tasks each to report. The cricket graphs look OK and the server page is all green in the relevant areas.

Any clue as to what may be happening? Did I offend the ETs or something?
____________

Profile Link
Avatar
Send message
Joined: 18 Sep 03
Posts: 827
Credit: 1,555,279
RAC: 551
Germany
Message 1301246 - Posted: 2 Nov 2012, 12:57:39 UTC - in response to Message 1301226.
Last modified: 2 Nov 2012, 12:59:00 UTC

Look in the panic thread, almost everyone has that problem right now. BTW, the cricket graphs looked OK few days ago, now the 100Mbit internet connection is maxed out / overloaded.
____________
.

jravin
Send message
Joined: 25 Mar 02
Posts: 929
Credit: 97,061,312
RAC: 86,632
United States
Message 1301263 - Posted: 2 Nov 2012, 13:59:35 UTC - in response to Message 1301246.

Look in the panic thread, almost everyone has that problem right now. BTW, the cricket graphs looked OK few days ago, now the 100Mbit internet connection is maxed out / overloaded.


Thanks. I'll check it out.
____________

KB7RZF
Volunteer tester
Avatar
Send message
Joined: 15 Aug 99
Posts: 9463
Credit: 3,085,541
RAC: 2,183
United States
Message 1302355 - Posted: 5 Nov 2012, 1:46:38 UTC

One trick that folks have been using is to set No New Tasks, click your update button, and then click to Allow New Tasks. That may or may not help, but just a thought.
____________

Profile Fred E.Project donor
Volunteer tester
Send message
Joined: 22 Jul 99
Posts: 768
Credit: 24,138,918
RAC: 383
United States
Message 1302360 - Posted: 5 Nov 2012, 2:05:44 UTC

One trick that folks have been using is to set No New Tasks, click your update button, and then click to Allow New Tasks. That may or may not help, but just a thought


yet both machines have nearly 400 tasks each to report

You definately need to go No New Tasks (NNT), and you'll have to do it multiple times because the server now accepts a max of 64 per update. When on NNT, you don't have to wait the 5 minutes between manual updates.


____________
Another Fred
Support SETI@home when you search the Web with GoodSearch or shop online with GoodShop.

Profile Brother Frank
Send message
Joined: 10 Dec 11
Posts: 26
Credit: 15,142,410
RAC: 0
United States
Message 1302375 - Posted: 5 Nov 2012, 3:19:51 UTC - in response to Message 1302360.
Last modified: 5 Nov 2012, 3:26:30 UTC

The technique you describe Fred of setting updates to no new tasks (NNT) and then updating has worked very well for me and the little crunchers around here. Sometimes I hit the do network communication too so that I'm able to break out of a long time out. Then when that clears the hour or so time out delay, I hit the manual update. It almost always works within a few minutes. After I get the update out, I set Boinc Manager to accept new work again. A few hours later I repeat the sequence again. I have noticed my core duo notebook and my i3 notebook run ok by themselves sometimes, but I'm always having to do this for my desktops and my new Dell XPS 15.6 i7-2670qm with a dedicated Nvidia GT 525m graphics card. I am wondering if the dedicated Nvidia graphics cards are part of this problem too because the computers without them are not having the update problem constantly in the last few days. I sure hope this can be addressed as part of maintenance next week or before. Please, Please. There is something else though that is bothering me a great deal more than this one issue though. It is a that there is far too much unreliability of the equipment and network that gets the signal data out of Seti to be processed by the volunteers and then back to Seti when it is processed. That unreliability is causing far too much wasted time and effort from a great number of people. It is probably getting to the point where it must be worked out and solved.

If this "burping and indigestion" needs a fund drive to solve it centered around new equipment, I would be happy to donate a hundred or two to be part of a solution with a funding drive of some kind when the folks back at Seti know what the equipment issues are. Really though, I don't know if this is a telecommunications bandwidth problem, a server and database problem, a power and voltage reliability problem, or some combination of the three. Could a technical expert in a set of complicated issues like this who has an interest in Seti help us get to the core problems and give us suggestions as a non-profit or even donate time to help us? I've just been a volunteer part of this for less than a year now, but would really like to see Seti@Home get to the bottom of this and solve it. It's about time to attack this issue. Brother Frank

Profile Fred E.Project donor
Volunteer tester
Send message
Joined: 22 Jul 99
Posts: 768
Credit: 24,138,918
RAC: 383
United States
Message 1302383 - Posted: 5 Nov 2012, 4:00:59 UTC

The technique you describe Fred of setting updates to no new tasks (NNT) and then updating has worked very well for me and the little crunchers around here. Sometimes I hit the do network communication too so that I'm able to break out of a long time out. Then when that clears the hour or so time out delay, I hit the manual update. It almost always works within a few minutes. After I get the update out, I set Boinc Manager to accept new work again. A few hours later I repeat the sequence again. I have noticed my core duo notebook and my i3 notebook run ok by themselves sometimes, but I'm always having to do this for my desktops and my new Dell XPS 15.6 i7-2670qm with a dedicated Nvidia GT 525m graphics card.

Good point - I also do this to reset the backoff time back to the project's 5 minutes.
am wondering if the dedicated Nvidia graphics cards are part of this problem too because the computers without them are not having the update problem constantly in the last few days.

Maybe - I have a working theory that the less results you are reporting, the better chance of getting a connect with a work request the past few days. My gpu almost always has a result every 5 minutes, so I've had a harder time than the cpu only machines.

I sure hope this can be addressed as part of maintenance next week or before. Please, Please. There is something else though that is bothering me a great deal more than this one issue though.
The staff intervened this afternoon and is definately aware of the problems. With the splitters shut down, I'm sure it will have top priority when they get in tomorrow.

There is something else though that is bothering me a great deal more than this one issue though. It is a that there is far too much unreliability of the equipment and network that gets the signal data out of Seti to be processed by the volunteers and then back to Seti when it is processed. That unreliability is causing far too much wasted time and effort from a great number of people. It is probably getting to the point where it must be worked out and solved.

If this "burping and indigestion" needs a fund drive to solve it centered around new equipment, I would be happy to donate a hundred or two to be part of a solution with a funding drive of some kind when the folks back at Seti know what the equipment issues are. Really though, I don't know if this is a telecommunications bandwidth problem, a server and database problem, a power and voltage reliability problem, or some combination of the three. Could a technical expert in a set of complicated issues like this who has an interest in Seti help us get to the core problems and give us suggestions as a non-profit or even donate time to help us? I've just been a volunteer part of this for less than a year now, but would really like to see Seti@Home get to the bottom of this and solve it. It's about time to attack this issue. Brother Frank

Can't argue with the unreliability point, it has bothered all of us. Steps are indeed underway. Slavac of the GPU Users Group formed a tax advantaged corp. to collect funds specifically for infra structure upgrades. It's set up to make sure the donations are deductible for US Taxpayers. To date, we've contibuted funds for 2 new servers that are now in service, as well as equipment for increased storage capacity. He's now awaiting specs from the project for another server. Watch his sticky thread in this forum for what's going on. He'd be happy to answer any questions you may have. Funding is a big issue - they recently cut the hours for a key staff person to half time due to funding shortfalls, and that has to be part of our current problems. Government and private grants for science projects are harder and harder to get, and the project is increasingly dependent on general purpose donations from crunchers as well as those designated for specific hardware. Hope all can participate in the current annnual fund drive.
____________
Another Fred
Support SETI@home when you search the Web with GoodSearch or shop online with GoodShop.

Message boards : Number crunching : Problems with Update

Copyright © 2014 University of California