Panic Mode On (109) Server Problems?

Message boards : Number crunching : Panic Mode On (109) Server Problems?
Message board moderation

To post messages, you must log in.

Previous · 1 . . . 34 · 35 · 36 · 37 · 38 · Next

AuthorMessage
Profile Wiggo "Socialist"
Avatar

Send message
Joined: 24 Jan 00
Posts: 13182
Credit: 175,542,370
RAC: 56,869
Australia
Message 1913638 - Posted: 17 Jan 2018, 23:42:36 UTC - in response to Message 1913634.  

I noticed that gbt_splitter#4 got briefly pulled from duty. Now back. I wonder if it was hanging up the output. Really frustrating that my Linux box is still unable to get any appreciable work while my slow Windows7 boxes have mostly full caches.

The guys are probably trying to break up the congregation of splitters working on the 1st GBT file in the list (still looks to be a half dozen splitters still working on the 1 file when you zoom in on it) and get them working on other files instead and you just caught them in the act. ;-)

Cheers.
ID: 1913638 · Report as offensive
Stephen "Heretic" Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 20 Sep 12
Posts: 2995
Credit: 57,679,704
RAC: 83,928
Australia
Message 1913642 - Posted: 17 Jan 2018, 23:59:39 UTC - in response to Message 1913634.  

I noticed that gbt_splitter#4 got briefly pulled from duty. Now back. I wonder if it was hanging up the output. Really frustrating that my Linux box is still unable to get any appreciable work while my slow Windows7 boxes have mostly full caches.


. . Ironic as it may be for once I have the opposite experience. Both Linux boxes received nearly full caches (over many requests) while the Windows box got work right away for the GPU but struggled over many many attempts before getting any work for the CPU. I had to reschedule WUs from the GPU Q.

Stephen

? ?
ID: 1913642 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 3558
Credit: 212,828,088
RAC: 294,890
United States
Message 1913643 - Posted: 18 Jan 2018, 0:02:35 UTC

I finally had to shut off gpu work entirely for the project and wait through 4 request cycles before the servers finally decided to send me 40 cpu tasks on the Linux machine. I haven't had any cpu tasks from Seti for 2 days now on that machine until just now.
Seti@Home classic workunits:20,676 CPU time:74,226 hours
ID: 1913643 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 3558
Credit: 212,828,088
RAC: 294,890
United States
Message 1913644 - Posted: 18 Jan 2018, 0:05:02 UTC - in response to Message 1913642.  

I noticed that gbt_splitter#4 got briefly pulled from duty. Now back. I wonder if it was hanging up the output. Really frustrating that my Linux box is still unable to get any appreciable work while my slow Windows7 boxes have mostly full caches.


. . Ironic as it may be for once I have the opposite experience. Both Linux boxes received nearly full caches (over many requests) while the Windows box got work right away for the GPU but struggled over many many attempts before getting any work for the CPU. I had to reschedule WUs from the GPU Q.

Stephen

? ?

I'm absolutely convinced that the project ignores Ryzen machines for work requests. Couldn't get work on the Windows Ryzen either.
Seti@Home classic workunits:20,676 CPU time:74,226 hours
ID: 1913644 · Report as offensive
Profile Chris Feyen
Volunteer tester

Send message
Joined: 1 Jan 01
Posts: 57
Credit: 8,866,429
RAC: 50,225
United States
Message 1913655 - Posted: 18 Jan 2018, 0:52:12 UTC

Creation time is at 72 a second at the moment. That's the highest I've ever seen it. Most of my machines have a full cache - a couple of them I did have to kick start though.

0 results ready to send and number of units received in the last hour is at 55k but I would imagine those numbers will return to normal in the next few hours.
~Chris

ID: 1913655 · Report as offensive
rob smith Special Project $250 donor
Volunteer tester

Send message
Joined: 7 Mar 03
Posts: 15649
Credit: 276,593,570
RAC: 235,947
United Kingdom
Message 1913685 - Posted: 18 Jan 2018, 6:27:06 UTC

Hard to say exactly what the state of play is with the replica being over an hour behind the master....
Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?
ID: 1913685 · Report as offensive
rob smith Special Project $250 donor
Volunteer tester

Send message
Joined: 7 Mar 03
Posts: 15649
Credit: 276,593,570
RAC: 235,947
United Kingdom
Message 1913695 - Posted: 18 Jan 2018, 10:35:09 UTC

....from the number of tasks on my crunchers it looks as if the splitters are spluttering again :-(
Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?
ID: 1913695 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 11738
Credit: 111,634,530
RAC: 48,595
United Kingdom
Message 1913700 - Posted: 18 Jan 2018, 11:13:31 UTC - in response to Message 1913695.  

I'd allowed one of my single GPU machines to help Einstein for a while. I've just given it a prod to restore normal service, and got the full 100 at the first request. As somebody said, it's the luck of the draw.
ID: 1913700 · Report as offensive
Profile JaundicedEye Project Donor
Avatar

Send message
Joined: 14 Mar 12
Posts: 3656
Credit: 28,670,356
RAC: 16,244
United States
Message 1913730 - Posted: 18 Jan 2018, 14:34:53 UTC

Caches are full and RAC is steadily dropping
............everything back to normal and the Credit Screw is turning as usual.

"Sour Grapes make a bitter Whine." <(0)>
ID: 1913730 · Report as offensive
Profile Mr. Kevvy Crowdfunding Project Donor*Special Project $250 donor
Volunteer moderator
Volunteer tester
Avatar

Send message
Joined: 15 May 99
Posts: 1859
Credit: 412,870,039
RAC: 364,003
Canada
Message 1913739 - Posted: 18 Jan 2018, 16:00:29 UTC - in response to Message 1913730.  

Caches are full and RAC is steadily dropping...


Indeed... the lack of work accelerated the fall to where it was going to end up anyways; ripping off the proverbial bandage.
“Never doubt that a small group of thoughtful, committed citizens can change the world; indeed, it's the only thing that ever has.”
--- Margaret Mead

ID: 1913739 · Report as offensive
Profile betreger Project Donor
Avatar

Send message
Joined: 29 Jun 99
Posts: 6977
Credit: 16,900,633
RAC: 12,898
United States
Message 1913743 - Posted: 18 Jan 2018, 16:30:25 UTC

Those who use Einstein as a backup project should take note and have a sufficient cache next Tuesday or they may very well run out of work.
Einstein has decided to join the fun by stating
We are going to shut down the project next Tuesday, Jan 23rd at around 10 AM CET for an upgrade of our database backend systems to make them ready for the years to come. We're going to upgrade hardware parts, operating systems as well the databases themselves, which is why we need to shut down the entire project, including the BOINC backend and this very website.

We should have the pleasure of a double outrage.
ID: 1913743 · Report as offensive
Profile Jimbocous Project Donor
Volunteer tester
Avatar

Send message
Joined: 1 Apr 13
Posts: 1062
Credit: 98,254,672
RAC: 86,628
United States
Message 1913778 - Posted: 18 Jan 2018, 18:42:44 UTC - in response to Message 1913743.  

Those who use Einstein as a backup project should take note and have a sufficient cache next Tuesday or they may very well run out of work.
Einstein has decided to join the fun by stating
We are going to shut down the project next Tuesday, Jan 23rd at around 10 AM CET for an upgrade of our database backend systems to make them ready for the years to come. We're going to upgrade hardware parts, operating systems as well the databases themselves, which is why we need to shut down the entire project, including the BOINC backend and this very website.

We should have the pleasure of a double outrage.

Yeah, it is a shame about the scheduling. Any other day would have sufficed ...
ID: 1913778 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 3558
Credit: 212,828,088
RAC: 294,890
United States
Message 1913807 - Posted: 18 Jan 2018, 20:58:17 UTC - in response to Message 1913778.  

Can always choose another backup project. I'll have MilkyWay and GPUGrid.net as backups also. Though if I build a big enough cache of Einstein work, that shouldn't be an issue either.
Seti@Home classic workunits:20,676 CPU time:74,226 hours
ID: 1913807 · Report as offensive
Profile betreger Project Donor
Avatar

Send message
Joined: 29 Jun 99
Posts: 6977
Credit: 16,900,633
RAC: 12,898
United States
Message 1913813 - Posted: 18 Jan 2018, 21:21:18 UTC - in response to Message 1913778.  

Yeah, it is a shame about the scheduling. Any other day would have sufficed ...

But a double outrage is something to behold.
ID: 1913813 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 11738
Credit: 111,634,530
RAC: 48,595
United Kingdom
Message 1913833 - Posted: 18 Jan 2018, 23:01:11 UTC - in response to Message 1913807.  

Can always choose another backup project. I'll have MilkyWay and GPUGrid.net as backups also. Though if I build a big enough cache of Einstein work, that shouldn't be an issue either.
Last time I crossed swords with MilkyWay, it was appallingly badly managed. And GPUGrid is giving me RSI in the mouse-click finger, because I run it but have extreme difficulty snagging new work.
ID: 1913833 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 3558
Credit: 212,828,088
RAC: 294,890
United States
Message 1913844 - Posted: 19 Jan 2018, 0:05:46 UTC - in response to Message 1913833.  

Can always choose another backup project. I'll have MilkyWay and GPUGrid.net as backups also. Though if I build a big enough cache of Einstein work, that shouldn't be an issue either.
Last time I crossed swords with MilkyWay, it was appallingly badly managed. And GPUGrid is giving me RSI in the mouse-click finger, because I run it but have extreme difficulty snagging new work.

MilkyWay?? Badly managed?? Wow, very different experience here.. MW is the most set and forget project I have run. I never have to micromanage it at all. I love the hard limit of 80 tasks per gpu at any one time. Never a chance of getting too much work and never any chance of running out. I only crunch gpu tasks so that means I have run the Binary Pulsar Search while it lasted and now run the Gamma Ray Pulsar Search. The only issues I have seen with the project is the occasional bad work unit that promptly gets tossed out very quickly. The servers seem to stay up for very long times, months at a time in fact.

Yes, I have just recently joined GPUGrid.net and the gpu work availability is very spotty and random. The tasks when they are made available are quickly gobbled up by many fingers bashing the update button. The cpu work for Linux hosts has had work available pretty much all the time. No problem getting cpu work for the Linux host. We are asking the project scientists to make the cpu work available for Windows too.
Seti@Home classic workunits:20,676 CPU time:74,226 hours
ID: 1913844 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 11738
Credit: 111,634,530
RAC: 48,595
United Kingdom
Message 1913850 - Posted: 19 Jan 2018, 0:27:43 UTC - in response to Message 1913844.  

Well, I found it necessary to make Post 58550. Read through to Post 58572, and note his titles.
ID: 1913850 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 3558
Credit: 212,828,088
RAC: 294,890
United States
Message 1913852 - Posted: 19 Jan 2018, 0:40:27 UTC - in response to Message 1913850.  
Last modified: 19 Jan 2018, 0:46:25 UTC

Well, I found it necessary to make Post 58550. Read through to Post 58572, and note his titles.

Well, even Project Scientists and Developers are human. Witness our own Eric K. and the recent spate of typo errors. I do remember they (MW) having issues initially with the n-body mt application but they sorted it out evidently and I didn't follow any of the threads since as I stated I don't do MW cpu work. I haven't seen many posts about n-body issues other than host configuration questions.

And the mt documentation must be mostly stable and understood by now as the mt cpu app deployed at GPUGrid just this month by a student had a relatively easy startup. It was nice to find it obeyed the app_config core usage setting so it didn't hog all cores on my Ryzen 1800X. I am using 4 cores to process the cpu tasks leaving the other cores for Seti cpu tasks.
Seti@Home classic workunits:20,676 CPU time:74,226 hours
ID: 1913852 · Report as offensive
juan BFP Special Project $75 donor
Volunteer tester
Avatar

Send message
Joined: 16 Mar 07
Posts: 6574
Credit: 364,486,076
RAC: 157,113
Panama
Message 1913859 - Posted: 19 Jan 2018, 1:09:44 UTC

MW is fine, but IIRC to do something productive there you need to have an AMD GPU, NV stuff has troubles to work with DP used by MW.

Does that changes?

GPUGrid with it's long time to crunch WU is not really a project to use as a backup. IMHO
ID: 1913859 · Report as offensive
Profile Jeff Buck Special Project $250 donor
Volunteer tester

Send message
Joined: 11 Feb 00
Posts: 1441
Credit: 148,707,946
RAC: 18,213
United States
Message 1913874 - Posted: 19 Jan 2018, 1:39:21 UTC - in response to Message 1913844.  

MW is the most set and forget project I have run. I never have to micromanage it at all.
I only keep a backup available on one of my crunch-only machines, just to make sure it maintains a little heat in the bedroom on chilly nights when SaH runs out of work. My first choice is Asteroids, but they're often out of work, too, so I added MilkyWay as a backup to the backup. The last time it ran on Windows was about 3 years ago. It worked fine. But that machine is now Linux, and when MilkyWay kicked in one night a couple months ago, it turned out to be a colossal waste of time. I don't remember how many tasks it ran, but when I checked the results the next day, I found that all but one of them had been marked Invalid. I think they all ran to completion without throwing any errors, but it was all just wasted electricity (except for the little bit of extra heat). I never did try to figure out what might have happened, just turned off MilkyWay and added Einstein for the next time that both SaH and Asteroids ran out.
ID: 1913874 · Report as offensive
Previous · 1 . . . 34 · 35 · 36 · 37 · 38 · Next

Message boards : Number crunching : Panic Mode On (109) Server Problems?


 
©2018 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.