Panic Mode On (37) Server problems


log in

Advanced search

Message boards : Number crunching : Panic Mode On (37) Server problems

1 · 2 · 3 · 4 . . . 12 · Next
Author Message
Profile arkaynProject donor
Volunteer tester
Avatar
Send message
Joined: 14 May 99
Posts: 3622
Credit: 48,535,582
RAC: 35,688
United States
Message 1022305 - Posted: 3 Aug 2010, 22:41:00 UTC

New one to continue on......
____________

Cosmic_Ocean
Avatar
Send message
Joined: 23 Dec 00
Posts: 2245
Credit: 8,588,060
RAC: 4,272
United States
Message 1022337 - Posted: 4 Aug 2010, 0:13:52 UTC

Did everyone get enough work?

Sadly, no. Last time I got work was before last week's shutdown. I had about 12 days of APs to chew on, but more than half of what I got were early-exits, so I burned through about a dozen in less than a day instead of four per day. Ran out last night, but the AP splitters were disabled all weekend.

Oh well. At least my room will be cooler for a couple of days.
____________

Linux laptop uptime: 1484d 22h 42m
Ended due to UPS failure, found 14 hours after the fact

ClaggyProject donor
Volunteer tester
Send message
Joined: 5 Jul 99
Posts: 4067
Credit: 32,877,487
RAC: 6,804
United Kingdom
Message 1022349 - Posted: 4 Aug 2010, 0:35:35 UTC
Last modified: 4 Aug 2010, 0:36:50 UTC

Did everyone get enough work?


I'm also a bit short of Astropulse work this week too, on both my T8100 and E8500, 9 and 7 unstarted tasks respectively, had lots of short tasks also,
both will be doing a bit more CPDN than normal, E8500 has plenty of Cuda work though, so it's throughput won't totally collapse,

Claggy

Dave
Avatar
Send message
Joined: 29 Mar 02
Posts: 774
Credit: 23,193,139
RAC: 0
United Kingdom
Message 1022489 - Posted: 4 Aug 2010, 14:30:59 UTC

Yes but laptop always seems a bit short & will barely make it with a few horsu to go. DCF jumped up to 0.9xxxxxx so I made it 0.09xxxxx...

Thinking about a router out of birthday money so my machines can b connected all the time so they never miss an opportunity to top-up (as I'm currently plugging/unplugging USB modem).

Profile James SotherdenProject donor
Avatar
Send message
Joined: 16 May 99
Posts: 8687
Credit: 33,073,167
RAC: 56,546
United States
Message 1022734 - Posted: 5 Aug 2010, 15:53:30 UTC
Last modified: 5 Aug 2010, 15:54:13 UTC

Well my i7 has plenty of GPU work, Friday I set no new work because I had a ton of it. Both GPU and CPU. The CPU ran out Wednesday morning. Still have 400 PLus GPU to crunch though. Im doing them 30 minutes each so still have 200 hours to go. Im worried though I see alot of Aug 13 times to report.
Im Wonedring if I should get new work friday.
The mac ran out today along with the old P4.

Seems like each outage has its own life dont it:)
____________

Old James

msattlerProject donor
Volunteer tester
Avatar
Send message
Joined: 9 Jul 00
Posts: 38870
Credit: 577,704,081
RAC: 524,150
United States
Message 1022736 - Posted: 5 Aug 2010, 15:56:33 UTC - in response to Message 1022734.

Well my i7 has plenty of GPU work, Friday I set no new work because I had a ton of it. Both GPU and CPU. The CPU ran out Wednesday morning. Still have 400 PLus GPU to crunch though. Im doing them 30 minutes each so still have 200 hours to go. Im worried though I see alot of Aug 13 times to report.
Im Wonedring if I should get new work friday.
The mac ran out today along with the old P4.

Seems like each outage has its own life dont it:)

Dunno why the GTS250 takes 30 minutes to do a WU, but I am not familiar with the speed of that card.

But you could use the rescheduler tool to shift some of those WUs over to the CPU and have it work on them rather than have it sit idle.......
____________
*********************************************
Embrace your inner kitty...ya know ya wanna!

I have met a few friends in my life.
Most were cats.

Profile James SotherdenProject donor
Avatar
Send message
Joined: 16 May 99
Posts: 8687
Credit: 33,073,167
RAC: 56,546
United States
Message 1022743 - Posted: 5 Aug 2010, 16:18:29 UTC

Thats my normal time with opp apps running a GPU task. Id thought of the reshedular, but I think this was a one time bloop. ( i hope ) So for now I will just let the GPU run.
____________

Old James

Profile [seti.international] Dirk SadowskiProject donor
Volunteer tester
Avatar
Send message
Joined: 6 Apr 07
Posts: 7060
Credit: 59,960,130
RAC: 21,761
Germany
Message 1022861 - Posted: 6 Aug 2010, 0:22:34 UTC - in response to Message 1022743.
Last modified: 6 Aug 2010, 0:29:14 UTC

Thats my normal time with opp apps running a GPU task. Id thought of the reshedular, but I think this was a one time bloop. ( i hope ) So for now I will just let the GPU run.


If you have bad luck and you get a resend VLAR WU for your GPU, the opt. CUDA VLARkill app will delete (kill) this VLAR WU. This will reset your daily WU quota of your GPU to 99. This means (this value x8) 792 WUs/GPU/day.
If you get a bunch of resend VLAR WUs, the daily WU quota will be much smaller.

O.K., you have 'only' a GTS250, but this GPU need also WUs. ;-)

I use Fred's BOINC Rescheduler to the time we'll see no longer resend VLAR WUs for GPU, which will be maybe still a few months.
____________
BR



>Das Deutsche Cafe. The German Cafe.<

Profile [seti.international] Dirk SadowskiProject donor
Volunteer tester
Avatar
Send message
Joined: 6 Apr 07
Posts: 7060
Credit: 59,960,130
RAC: 21,761
Germany
Message 1022863 - Posted: 6 Aug 2010, 0:26:46 UTC

Hey, what's going on?? My BOINC can't UL/DL/report!!


:o)


Couldn't resist.. X-D

____________
BR



>Das Deutsche Cafe. The German Cafe.<

ratidhux
Send message
Joined: 8 Apr 00
Posts: 3
Credit: 1,753,740
RAC: 3,324
United Kingdom
Message 1022948 - Posted: 6 Aug 2010, 7:09:22 UTC

I now have 44 "ghost" tasks attributed to my computer. They started to appear after the introduction of the new 3 day outage regime. 31 after the first outage, 1 after the second and 13 last week. All 44 still appear on the "In progress" area of the "Tasks" section of my personal account on the web site. However, none of them have ever downloaded to my home computer for processing. Is this uniquely my problem and/or am I doing something wrong that is producing this anomally? I would welcome an explanation.
____________

Profile soft^spirit
Avatar
Send message
Joined: 18 May 99
Posts: 6374
Credit: 28,631,059
RAC: 94
United States
Message 1022955 - Posted: 6 Aug 2010, 7:29:50 UTC

1 GPU dry.
1 CPU should go dry in about 6 hours
1 GPU has about 36 hours left..
1 CPU still busy on CPDN. Should start trying to get SETI work sometime this
weekend.


____________

Janice

ratidhux
Send message
Joined: 8 Apr 00
Posts: 3
Credit: 1,753,740
RAC: 3,324
United Kingdom
Message 1022962 - Posted: 6 Aug 2010, 7:58:30 UTC

One was timed out and now appears in in the "Error" section = 45. Just in case agitation is caused by the original posting not adding up to 45.

P.S. A big thanks to all the staff at SETI for keeping it all together in the face of persistant adversity.
____________

Profile [seti.international] Dirk SadowskiProject donor
Volunteer tester
Avatar
Send message
Joined: 6 Apr 07
Posts: 7060
Credit: 59,960,130
RAC: 21,761
Germany
Message 1023000 - Posted: 6 Aug 2010, 9:57:16 UTC - in response to Message 1022962.

After ~ 10 years membership.. Welcome at the forum!

You can't do anything for to protect your BOINC for ghost WUs.
If the traffic of the SETI@home internet access/cable is maxed out (cricket graph, in/what means DL to us, ~ 95 Mbits/s) and your BOINC ask for new work, it can happen that you get ghost WUs.

____________
BR



>Das Deutsche Cafe. The German Cafe.<

ratidhux
Send message
Joined: 8 Apr 00
Posts: 3
Credit: 1,753,740
RAC: 3,324
United Kingdom
Message 1023009 - Posted: 6 Aug 2010, 11:10:00 UTC

Feedback much appreciated. Thanks.
____________

Profile SciManStevProject donor
Volunteer tester
Avatar
Send message
Joined: 20 Jun 99
Posts: 4824
Credit: 80,926,224
RAC: 38,025
United States
Message 1023044 - Posted: 6 Aug 2010, 14:01:24 UTC

My rig ran dry yesterday morning, but on the good side once I upload and report I will be able to de-attach and re-attach to dump my ghost load. I had been wanting to for the last several outages, but I still had work and wouldn't do it. It is harder to reload suffeciently from a cold start, but with the way the limits have evolved, it should be much easier.

Steve
____________
Warning, addicted to SETI crunching!
Crunching as a member of GPU Users Group.
GPUUG Website

Bernd Noessler
Send message
Joined: 15 Nov 09
Posts: 99
Credit: 52,635,434
RAC: 0
Germany
Message 1023056 - Posted: 6 Aug 2010, 14:17:13 UTC

A hint for the developers:
I have 7 Linux boxes running Seti.
Until now I had not one ghost attached to them.
For me it looks like a "Windows only" problem.

Profile soft^spirit
Avatar
Send message
Joined: 18 May 99
Posts: 6374
Credit: 28,631,059
RAC: 94
United States
Message 1023059 - Posted: 6 Aug 2010, 14:20:27 UTC

19 minutes until about 10 tasks are past the deadline... come up uploads!!

____________

Janice

Profile SciManStevProject donor
Volunteer tester
Avatar
Send message
Joined: 20 Jun 99
Posts: 4824
Credit: 80,926,224
RAC: 38,025
United States
Message 1023063 - Posted: 6 Aug 2010, 14:25:07 UTC - in response to Message 1023056.

A hint for the developers:
I have 7 Linux boxes running Seti.
Until now I had not one ghost attached to them.
For me it looks like a "Windows only" problem.


I think it may also have something to do with how much work each machine has to download. My one rig has to download well over 2000 wu's to sustain it for 3 days. I often have to battle maxed out bandwidth, and continously babysit my rig to get enough work, which really doesn't happen until the limits are taken off. If I needed much fewer wu's, then it would fill up much faster, hence less chance of ghosts. Just my thoughts, and I could be completely wrong. :)

Steve
____________
Warning, addicted to SETI crunching!
Crunching as a member of GPU Users Group.
GPUUG Website

Profile hiamps
Volunteer tester
Avatar
Send message
Joined: 23 May 99
Posts: 4292
Credit: 72,971,319
RAC: 0
United States
Message 1023064 - Posted: 6 Aug 2010, 14:27:53 UTC - in response to Message 1023059.

19 minutes until about 10 tasks are past the deadline... come up uploads!!

Seems uploads were going by this time last week....
____________
Official Abuser of Boinc Buttons...
And no good credit hound!

Josef W. SegurProject donor
Volunteer developer
Volunteer tester
Send message
Joined: 30 Oct 99
Posts: 4227
Credit: 1,042,553
RAC: 346
United States
Message 1023067 - Posted: 6 Aug 2010, 14:31:58 UTC - in response to Message 1023056.

A hint for the developers:
I have 7 Linux boxes running Seti.
Until now I had not one ghost attached to them.
For me it looks like a "Windows only" problem.

LOL, I think it is a server-side issue almost exclusively, which would make it "Linux only". In reality, it is of course just a weakness of the BOINC system and some programming effort is needed to make it more robust.
Joe

1 · 2 · 3 · 4 . . . 12 · Next

Message boards : Number crunching : Panic Mode On (37) Server problems

Copyright © 2014 University of California