Manually uploading client_state.xml to s@h


log in

Advanced search

Message boards : Number crunching : Manually uploading client_state.xml to s@h

1 · 2 · 3 · Next
Author Message
Profile -= Vyper =-Project donor
Volunteer tester
Avatar
Send message
Joined: 5 Sep 99
Posts: 1092
Credit: 326,048,604
RAC: 67,473
Sweden
Message 1033608 - Posted: 17 Sep 2010, 16:59:57 UTC
Last modified: 17 Sep 2010, 17:00:12 UTC

When the s@h client has troubles getting through to s@h when trying to upload the file to report completed workunits, is there an manual option that you could do to make it work?
That file is rapidly growing and in my case that file has become nearly 60 Mbyte in size and i truly understand that there are troubles getting it to upload completely.
Is there a way to do it or should i hope that i don't have crunched more than a week worth of WU's which can't be reported?
Suggestions anyone even if it's bad news!

Kind regards Vyper
____________

_________________________________________________________________________
Addicted to SETI crunching!
Founder of GPU Users Group

Profile -= Vyper =-Project donor
Volunteer tester
Avatar
Send message
Joined: 5 Sep 99
Posts: 1092
Credit: 326,048,604
RAC: 67,473
Sweden
Message 1033614 - Posted: 17 Sep 2010, 17:36:30 UTC - in response to Message 1033609.


Good question to ask......
I was thinking exactly the same thing yesterday when I was getting nothing but HTTP internal server errors every time my rigs tried to report their stored up completed work.
Thankfully, things settled down a bit and they finally all made it through. Don't know the size of the file, but I think the largest had about 3000 WUs to report. Yours must have a lot more than that.

I too, was wondering if there was a way to get Boinc to report only X number of completed WUs at a time, or be able to split the file up into smaller subfiles to transfer successfully.

I think the Boinc devs should consider adding an option.....
<report_max_results>XXXXX</report_max_results>


Yea i wrote that long ago that it should be split and that it would be incorporated in the Boinc structure itself not only to s@h.
Small batches would be better for the overall amount of TCP packets needed for resend etc.
For now that machine has about 4800 Wu's to report and it's increasing fast :-/
Hope there is a way otherwise i need to babysit and watch the graphs so the network isn't congested to it's max and wait it out and hope :)

Kind regards Vyper
____________

_________________________________________________________________________
Addicted to SETI crunching!
Founder of GPU Users Group

Profile Ageless
Avatar
Send message
Joined: 9 Jun 99
Posts: 12399
Credit: 2,669,398
RAC: 850
Netherlands
Message 1033694 - Posted: 17 Sep 2010, 21:28:36 UTC

It isn't client_state.xml that you're trying to upload. It's sched_request_*.xml, so you have to check the size of this file.

What error message do you get eventually? A time out?
____________
Jord

Fighting for the correct use of the apostrophe, together with Weird Al Yankovic

Profile -= Vyper =-Project donor
Volunteer tester
Avatar
Send message
Joined: 5 Sep 99
Posts: 1092
Credit: 326,048,604
RAC: 67,473
Sweden
Message 1034069 - Posted: 18 Sep 2010, 16:04:39 UTC - in response to Message 1034005.
Last modified: 18 Sep 2010, 16:48:41 UTC

Gonna try now! As i write this.
____________

_________________________________________________________________________
Addicted to SETI crunching!
Founder of GPU Users Group

Profile -= Vyper =-Project donor
Volunteer tester
Avatar
Send message
Joined: 5 Sep 99
Posts: 1092
Credit: 326,048,604
RAC: 67,473
Sweden
Message 1034090 - Posted: 18 Sep 2010, 16:50:15 UTC - in response to Message 1034072.
Last modified: 18 Sep 2010, 16:52:27 UTC

Boinc has behaved badly so the machine was set in an unresponsive state tonight so it hasn't been up until my colleague rebooted it at the office this afternoon (4 hours ago)..
So i believe it's about 5400 wu's which need to be reported. First i need to get rid of all the WU which can't be downloaded in the queue so i'm aborting them..

Otherwise boinc can't work properly and starts to loose track of heartbeats to the clients putting the machine in a extreme unresponsive state when remotecontrolling it.

Regards Vyper
____________

_________________________________________________________________________
Addicted to SETI crunching!
Founder of GPU Users Group

Profile -= Vyper =-Project donor
Volunteer tester
Avatar
Send message
Joined: 5 Sep 99
Posts: 1092
Credit: 326,048,604
RAC: 67,473
Sweden
Message 1034113 - Posted: 18 Sep 2010, 17:50:33 UTC
Last modified: 18 Sep 2010, 17:58:19 UTC

Ok

Sched request is up to 31 MB in size and is about to report 8448 tasks.
Wish me luck :D

Kind regards Vyper
____________

_________________________________________________________________________
Addicted to SETI crunching!
Founder of GPU Users Group

Profile perryjay
Volunteer tester
Avatar
Send message
Joined: 20 Aug 02
Posts: 3377
Credit: 16,212,599
RAC: 4,131
United States
Message 1034123 - Posted: 18 Sep 2010, 18:07:25 UTC

Somebody must be getting through, my total and RAC are climbing slowly and my pendings are dropping. Good luck Vyper.
____________


PROUD MEMBER OF Team Starfire World BOINC

Profile -= Vyper =-Project donor
Volunteer tester
Avatar
Send message
Joined: 5 Sep 99
Posts: 1092
Credit: 326,048,604
RAC: 67,473
Sweden
Message 1034127 - Posted: 18 Sep 2010, 18:10:34 UTC
Last modified: 18 Sep 2010, 18:11:38 UTC

Ok the first try didn't work.

It tried for about 20 minutes, then i got HTTP internal server error..
Lets try it again.. 30 MB here we go..

Pinching some needles in a voodoo doll..

Need some more "meowa meowa kazzzam." :D


//Vyper
____________

_________________________________________________________________________
Addicted to SETI crunching!
Founder of GPU Users Group

Profile Ageless
Avatar
Send message
Joined: 9 Jun 99
Posts: 12399
Credit: 2,669,398
RAC: 850
Netherlands
Message 1034128 - Posted: 18 Sep 2010, 18:11:08 UTC - in response to Message 1034113.
Last modified: 18 Sep 2010, 18:12:17 UTC

/me crosses everything crossable. :D

(Too late)

(I am emailing this all to David)
____________
Jord

Fighting for the correct use of the apostrophe, together with Weird Al Yankovic

Profile Ageless
Avatar
Send message
Joined: 9 Jun 99
Posts: 12399
Credit: 2,669,398
RAC: 850
Netherlands
Message 1034143 - Posted: 18 Sep 2010, 18:38:32 UTC

Vyper, David asks if you can email him your sched_request*.xml file. (I'll ask you to zip or otherwise compress it first. :-))


____________
Jord

Fighting for the correct use of the apostrophe, together with Weird Al Yankovic

Profile Ageless
Avatar
Send message
Joined: 9 Jun 99
Posts: 12399
Credit: 2,669,398
RAC: 850
Netherlands
Message 1034146 - Posted: 18 Sep 2010, 18:45:02 UTC

Update.

David Anderson wrote:
I see what the problem is (it's in the server; it crashes when lots of results are reported). I'll try to get it fixed today.

____________
Jord

Fighting for the correct use of the apostrophe, together with Weird Al Yankovic

Profile perryjay
Volunteer tester
Avatar
Send message
Joined: 20 Aug 02
Posts: 3377
Credit: 16,212,599
RAC: 4,131
United States
Message 1034150 - Posted: 18 Sep 2010, 18:50:12 UTC - in response to Message 1034146.

Great, that should be a load off the minds of our super-crunchers out there.
____________


PROUD MEMBER OF Team Starfire World BOINC

Sten-Arne
Volunteer tester
Send message
Joined: 1 Nov 08
Posts: 3692
Credit: 21,181,726
RAC: 4,893
Sweden
Message 1034152 - Posted: 18 Sep 2010, 18:53:44 UTC - in response to Message 1034150.

Great, that should be a load off the minds of our super-crunchers out there.


Well, it sure doesn't bother us "peons" anyhow :-)
____________

Profile [seti.international] Dirk Sadowski
Volunteer tester
Avatar
Send message
Joined: 6 Apr 07
Posts: 7117
Credit: 61,264,148
RAC: 4,623
Germany
Message 1034154 - Posted: 18 Sep 2010, 18:58:33 UTC - in response to Message 1034146.
Last modified: 18 Sep 2010, 19:01:02 UTC

Update.

David Anderson wrote:
I see what the problem is (it's in the server; it crashes when lots of results are reported). I'll try to get it fixed today.


Thanks!

In past I lost a few times a lot of results/Cr., because I needed to detach/attach that my GPU machine got again well contact to the server..

Currently 'top_host'_#4 owner.. ;-)
____________
BR

SETI@home Needs your Help ... $10 & U get a Star!

Team seti.international

Das Deutsche Cafe. The German Cafe.

1 · 2 · 3 · Next

Message boards : Number crunching : Manually uploading client_state.xml to s@h

Copyright © 2014 University of California