Panic Mode On (83) Server Problems?


log in

Advanced search

Message boards : Number crunching : Panic Mode On (83) Server Problems?

Previous · 1 . . . 13 · 14 · 15 · 16 · 17 · 18 · 19 . . . 22 · Next
Author Message
Grant (SSSF)
Send message
Joined: 19 Aug 99
Posts: 5864
Credit: 60,572,158
RAC: 47,668
Australia
Message 1365001 - Posted: 6 May 2013, 1:03:37 UTC - in response to Message 1364726.


Hmm, there was a spike in downloads, then it dropped down consideraly. About the same time i got a whole bunch of donwloads that wouldn't.
They've only just downloaded now.
____________
Grant
Darwin NT.

Profile HAL9000
Volunteer tester
Avatar
Send message
Joined: 11 Sep 99
Posts: 4440
Credit: 119,050,671
RAC: 139,982
United States
Message 1365119 - Posted: 6 May 2013, 13:23:55 UTC

Sorry guys. I'm getting all the WU's right now.
5/6/2013 9:00:12 AM SETI@home Scheduler request completed: got 79 new tasks

That was one of my machines first request after getting "Not requesting tasks: project is not highest priority" all weekend and running down its queue. I have to saw BOINV 7 can really be pants about work fetch. Giving it didn't even check the 2nd project since the 30th of last month.
____________
SETI@home classic workunits: 93,865 CPU time: 863,447 hours

Join the BP6/VP6 User Group today!

N9JFE David SProject donor
Volunteer tester
Avatar
Send message
Joined: 4 Oct 99
Posts: 11983
Credit: 14,652,494
RAC: 12,474
United States
Message 1365127 - Posted: 6 May 2013, 14:38:28 UTC

Good gravy, there's a lot of b*tching in here about not getting enough work.

I don't pay close attention to what I have on hand, usually just checking once a day on weekdays. Occasionally, my i7 has a couple fewer in progress than the limit, but it has not come close to running out since the server move outage (if then). Actually, it has let itself run dry of Einstein for now. (Or is Einstein also having feed problems and I haven't noticed?)

I forget, are the colo guys supposed to be loading splitter tapes (drives), or do the project guys still have to do that? If it's only the project guys, well, hey, it was the weekend. They only go in on their own time to fix major problems, and major problems are what the colo guys ARE there to take care of.

____________
David
Sitting on my butt while others boldly go,
Waiting for a message from a small furry creature from Alpha Centauri.


Profile Donald L. JohnsonProject donor
Avatar
Send message
Joined: 5 Aug 02
Posts: 6261
Credit: 738,877
RAC: 1,170
United States
Message 1365193 - Posted: 6 May 2013, 17:28:50 UTC - in response to Message 1365143.

I forget, are the colo guys supposed to be loading splitter tapes (drives), or do the project guys still have to do that? If it's only the project guys, well, hey, it was the weekend. They only go in on their own time to fix major problems, and major problems are what the colo guys ARE there to take care of.

As far as most of us are guessing, and I don't think the actual process has been confirmed by Matt or Eric, the hard drives from Arecibo still go to the Seti lab.
And the data is then transferred over the network link to the splitters at the colo. This we suspect is what causes the huge spikes in the upload link on the cricket graphs (the blue line, which is inbound to the servers).
How much of this process is automated and how much requires manual intervention by the boyz in da lab is up to guesses. Perhaps Matt can clarify some day.

As I recall from Matt's posts, the data from Arecibo is downloaded from the transfer drives at the lab, then sent to archive storage. As the supply of ready-to-split data files gets low, new data files are called up from storage, processed for RFI blanking, then transferred down the hill to the splitter storage portion of the JBOD (Or the RFI blanking may be done by a background server that was also moved to the CoLo). These transfers are done manually by Matt/Jeff/etc.

Once the ready-to-split files are transferred to the JBOD, when the splitters get to a certain low point, more files are automatically mounted for splitting.
____________
Donald
Infernal Optimist / Submariner, retired

Sten-Arne
Volunteer tester
Send message
Joined: 1 Nov 08
Posts: 3547
Credit: 20,740,790
RAC: 23,351
Sweden
Message 1365240 - Posted: 6 May 2013, 19:10:18 UTC
Last modified: 6 May 2013, 19:11:12 UTC

Heh, no benefits with a reasonably fast GPU (ATI HD7870) when almost all AP's I received are highly blanked 50-75%. Most of the work are done on the CPU, and that's an old Q8200 :-)

Yeah well, perhaps somewhere in the next batch of AP's, there might be more low blanked units.
____________

musicplayer
Send message
Joined: 17 May 10
Posts: 1457
Credit: 712,705
RAC: 981
Message 1365249 - Posted: 6 May 2013, 19:32:34 UTC
Last modified: 6 May 2013, 19:35:54 UTC

Blanked? Do you perhaps mean RFI?

Anyway, this perhaps should be put up somewhere else here, but the four types of numbers the Seti@home client is supposed to catch may not necessarily be what may be in an alien signal.

Therefore it is likely to go undetected.

Regardless of a task being a shorty or not (carrying out the gaussian search in the second instance), the numbers returned for spike, pulse and possible triplet should be the same otherwise.

Still, many people are convinced that narrowband signals do not consist of gaussians at all, whether or not they are computed by means of the spikes and pulses which may be present in a given task.

In other words, there should exists an algorithm which filters out all the numbers which may be the result of natural causes in order to be able to catch the signal (or signal curve) which may be present in an extraterrestrial signal.

However it may not be that interesting to process data sets which may be mostly empty when it comes to relevant contents.

Profile Donald L. JohnsonProject donor
Avatar
Send message
Joined: 5 Aug 02
Posts: 6261
Credit: 738,877
RAC: 1,170
United States
Message 1365345 - Posted: 7 May 2013, 5:32:50 UTC - in response to Message 1365249.

Blanked? Do you perhaps mean RFI?

Anyway, this perhaps should be put up somewhere else here

Indeed, probably the Seti Science forum.....
____________
Donald
Infernal Optimist / Submariner, retired

N9JFE David SProject donor
Volunteer tester
Avatar
Send message
Joined: 4 Oct 99
Posts: 11983
Credit: 14,652,494
RAC: 12,474
United States
Message 1365423 - Posted: 7 May 2013, 13:43:35 UTC

Okay, panickers, what say you now?

MB ready to send is now upward of 245,000, still with only 6 splitters running. (FWIW, AP ready to send is about 2,000 with all its splitters running.)

Can we still say the MB/PFB splitters can't keep up with demand? Or is something else going on that's been causing the ready number to be so low recently?

____________
David
Sitting on my butt while others boldly go,
Waiting for a message from a small furry creature from Alpha Centauri.


Profile Fred E.Project donor
Volunteer tester
Send message
Joined: 22 Jul 99
Posts: 768
Credit: 24,139,004
RAC: 1
United States
Message 1365429 - Posted: 7 May 2013, 14:01:06 UTC
Last modified: 7 May 2013, 14:01:49 UTC

Okay, panickers, what say you now?

MB ready to send is now upward of 245,000, still with only 6 splitters running. (FWIW, AP ready to send is about 2,000 with all its splitters running.)

Can we still say the MB/PFB splitters can't keep up with demand? Or is something else going on that's been causing the ready number to be so low recently?

Not convinced. They had real problems when (a) no AP was available, and (b) the MB work was mostly shorties. They are keeping up now with a better ratio of long/short MB tasks and the AP tasks. For example, I've received 364 tasks in the last 24 hours per BOINCTasks (staying at limits for the period). That number was ranging from 700 to 800 when we got almost all shorties, no AP, and the ready to send was depleted. Maybe there was a fix, but I'll wait and see.
____________
Another Fred
Support SETI@home when you search the Web with GoodSearch or shop online with GoodShop.

Profile HAL9000
Volunteer tester
Avatar
Send message
Joined: 11 Sep 99
Posts: 4440
Credit: 119,050,671
RAC: 139,982
United States
Message 1365437 - Posted: 7 May 2013, 14:14:19 UTC - in response to Message 1365429.
Last modified: 7 May 2013, 14:15:57 UTC

Okay, panickers, what say you now?

MB ready to send is now upward of 245,000, still with only 6 splitters running. (FWIW, AP ready to send is about 2,000 with all its splitters running.)

Can we still say the MB/PFB splitters can't keep up with demand? Or is something else going on that's been causing the ready number to be so low recently?

Not convinced. They had real problems when (a) no AP was available, and (b) the MB work was mostly shorties. They are keeping up now with a better ratio of long/short MB tasks and the AP tasks. For example, I've received 364 tasks in the last 24 hours per BOINCTasks (staying at limits for the period). That number was ranging from 700 to 800 when we got almost all shorties, no AP, and the ready to send was depleted. Maybe there was a fix, but I'll wait and see.

I am wondering if they were perhaps stress testing the new pfb splitters. Then again it could have been since the number of results & wu's to purge had reached such a large number the creation rate was being limited to prevent the db choking again.
____________
SETI@home classic workunits: 93,865 CPU time: 863,447 hours

Join the BP6/VP6 User Group today!

rob smithProject donor
Volunteer tester
Send message
Joined: 7 Mar 03
Posts: 8535
Credit: 59,378,342
RAC: 84,329
United Kingdom
Message 1365522 - Posted: 7 May 2013, 20:45:46 UTC

Hmmm, strange
I have two computers, both running Windows 7 sharing the same internet connection. Since today's outage one has been giving "Scheduler request failed: HTTP internal server error" every time it tries to connect, the other one connects, uploads, downloads and reports very smoothly.
The only difference I can see is the one that fails is trying to report over 100 tasks, while the one that succeeds is doing a handful at a time....
____________
Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?

Profile Fred E.Project donor
Volunteer tester
Send message
Joined: 22 Jul 99
Posts: 768
Credit: 24,139,004
RAC: 1
United States
Message 1365530 - Posted: 7 May 2013, 21:03:38 UTC

Hmmm, strange
I have two computers, both running Windows 7 sharing the same internet connection. Since today's outage one has been giving "Scheduler request failed: HTTP internal server error" every time it tries to connect, the other one connects, uploads, downloads and reports very smoothly.
The only difference I can see is the one that fails is trying to report over 100 tasks, while the one that succeeds is doing a handful at a time....

I connected once (no tasks) and haven't connected since. Also I can't get to my tasks page. Cricket says all is not well.
____________
Another Fred
Support SETI@home when you search the Web with GoodSearch or shop online with GoodShop.

Profile Gary CharpentierProject donor
Volunteer tester
Avatar
Send message
Joined: 25 Dec 00
Posts: 12734
Credit: 7,269,747
RAC: 17,656
United States
Message 1365531 - Posted: 7 May 2013, 21:04:27 UTC

I'm getting can't connect to server. Looked at the cricket and it looks like someone put a 100meg limit on it!

____________

Lionel
Send message
Joined: 25 Mar 00
Posts: 576
Credit: 235,858,758
RAC: 228,349
Australia
Message 1365539 - Posted: 7 May 2013, 21:29:40 UTC - in response to Message 1365522.

I have two computers, both running Windows 7 sharing the same internet connection. Since today's outage one has been giving "Scheduler request failed: HTTP internal server error" every time it tries to connect....


All of mine are doing the same. All have over 100 tasks to report.



____________

Profile MikeProject donor
Volunteer tester
Avatar
Send message
Joined: 17 Feb 01
Posts: 24537
Credit: 33,864,086
RAC: 23,615
Germany
Message 1365541 - Posted: 7 May 2013, 21:31:42 UTC

Reports are O.K. here but dont get new work.

____________

Lionel
Send message
Joined: 25 Mar 00
Posts: 576
Credit: 235,858,758
RAC: 228,349
Australia
Message 1365544 - Posted: 7 May 2013, 21:43:15 UTC - in response to Message 1365541.


And now the bun fight has begun ...



____________

Profile Wiggo
Avatar
Send message
Joined: 24 Jan 00
Posts: 7351
Credit: 96,866,163
RAC: 66,729
Australia
Message 1365546 - Posted: 7 May 2013, 22:07:09 UTC - in response to Message 1365544.

Things are starting to move here now.

Cheers.

Profile Gary CharpentierProject donor
Volunteer tester
Avatar
Send message
Joined: 25 Dec 00
Posts: 12734
Credit: 7,269,747
RAC: 17,656
United States
Message 1365552 - Posted: 7 May 2013, 22:50:14 UTC

Looks like the limit came off the link and we peaked at a hair over 500 meg.

____________

Previous · 1 . . . 13 · 14 · 15 · 16 · 17 · 18 · 19 . . . 22 · Next

Message boards : Number crunching : Panic Mode On (83) Server Problems?

Copyright © 2014 University of California