Experiment for server operations check...

Message boards : News : Experiment for server operations check...
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4 · 5 · Next

AuthorMessage
Christoph
Volunteer tester

Send message
Joined: 16 Oct 09
Posts: 58
Credit: 662,990
RAC: 0
Germany
Message 43754 - Posted: 22 Sep 2012, 21:08:28 UTC

Eric, one thing for you to think about: If you know some datafiles which are clean (not noisy or otherwise outliers) maybe you should keep these on hand and use always when you want/need to get things like runtimes and credit estimates straight quickly.

Since these test runs are not going inside the master science database anyway it doesn't matter if we get 'recycled' tasks, right?

At least I wouldn't complain if I have to crunch always the same datafiles every couple of month if things can be speed up by this.
Christoph
ID: 43754 · Report as offensive
zombie67 [MM]
Volunteer tester
Avatar

Send message
Joined: 18 May 06
Posts: 280
Credit: 26,477,429
RAC: 144
United States
Message 43755 - Posted: 22 Sep 2012, 22:58:39 UTC

Maybe this has nothing to do with the testing, but uploads are getting stuck. This started sometime within the last 24 hours.
Dublin, California
Team: SETI.USA

ID: 43755 · Report as offensive
Profile Eric J Korpela
Volunteer moderator
Project administrator
Project developer
Project scientist
Avatar

Send message
Joined: 15 Mar 05
Posts: 1547
Credit: 27,028,254
RAC: 1,959
United States
Message 43756 - Posted: 22 Sep 2012, 23:36:42 UTC - in response to Message 43755.  

Not seeing any indication of problems in the logs on this end.
ID: 43756 · Report as offensive
Richard Haselgrove
Volunteer tester

Send message
Joined: 3 Jan 07
Posts: 1451
Credit: 3,268,851
RAC: 208
United Kingdom
Message 43757 - Posted: 22 Sep 2012, 23:47:45 UTC - in response to Message 43756.  

Not seeing any indication of problems in the logs on this end.

May not be in the logs, but have a look at cricket. Data flows have been erratic all day: much chatter on the main project message board.
ID: 43757 · Report as offensive
Father Ambrose
Volunteer tester

Send message
Joined: 1 May 07
Posts: 556
Credit: 6,447,316
RAC: 281
United Kingdom
Message 43758 - Posted: 23 Sep 2012, 0:01:35 UTC

All ATI WU’s failed over night 197 (0xc5) EXIT_TIME_LIMIT_EXCEEDED



wuid=11161611


Michael
ID: 43758 · Report as offensive
Profile Eric J Korpela
Volunteer moderator
Project administrator
Project developer
Project scientist
Avatar

Send message
Joined: 15 Mar 05
Posts: 1547
Credit: 27,028,254
RAC: 1,959
United States
Message 43759 - Posted: 23 Sep 2012, 0:12:59 UTC - in response to Message 43758.  

That (the over limit problem) should fix itself over time.
ID: 43759 · Report as offensive
zombie67 [MM]
Volunteer tester
Avatar

Send message
Joined: 18 May 06
Posts: 280
Credit: 26,477,429
RAC: 144
United States
Message 43760 - Posted: 23 Sep 2012, 0:27:35 UTC - in response to Message 43757.  

Not seeing any indication of problems in the logs on this end.

May not be in the logs, but have a look at cricket. Data flows have been erratic all day: much chatter on the main project message board.



Yeah. The failures to upload are "http transient errors".
Dublin, California
Team: SETI.USA

ID: 43760 · Report as offensive
Claggy
Volunteer tester

Send message
Joined: 29 May 06
Posts: 1037
Credit: 8,440,339
RAC: 0
United Kingdom
Message 43764 - Posted: 23 Sep 2012, 8:57:02 UTC - in response to Message 43759.  

That (the over limit problem) should fix itself over time.

How soon? ihatelolcats's host is failing on both opencl_ati_100 & ati_opencl_100 plan classes, he's trying to do an app_info to run annoymous platform, i fear once he's got that right his Wu's will still error out,

Claggy
ID: 43764 · Report as offensive
Christoph
Volunteer tester

Send message
Joined: 16 Oct 09
Posts: 58
Credit: 662,990
RAC: 0
Germany
Message 43765 - Posted: 23 Sep 2012, 9:18:25 UTC - in response to Message 43764.  
Last modified: 23 Sep 2012, 9:21:35 UTC

To stop the error it is necessary to edit the client_state file.
But if the user have already much trouble with correcting the app_info maybe it is no good idea to point him there.

EDIT: in case you want to point to editing the client_state see here: http://setiweb.ssl.berkeley.edu/beta/forum_thread.php?id=1941&postid=43454#43454
Christoph
ID: 43765 · Report as offensive
Profile Sutaru Tsureku
Volunteer tester

Send message
Joined: 7 Jun 09
Posts: 285
Credit: 2,822,466
RAC: 0
Germany
Message 43766 - Posted: 23 Sep 2012, 14:59:05 UTC
Last modified: 23 Sep 2012, 15:06:00 UTC

I attached my machine new to SETI@home Beta (hostid=60271).

At the beginning the estimated times was ~ 14 hours for a SETI@home v7 WU.

Now after 6 granted and ~ 10 pending results, the estimated times are now at ~ 4 hours, so how the real calculate time is.

Credits are 100.99 to 140.75 for a SETI@home v7 result.


By the way ..
My machine have problems to upload and download at SETI@home and -Beta.
A few retries are needed for complete.

Again VHAR storm and clogged internet pipe?
.. because currently my machine get nearly only VHAR WUs at SETI@home.


- Best regards! :-) - Sutaru Tsureku, team seti.international founder. - Optimize your PC for higher RAC (@ SETI@home Main). - SETI@home needs your help. -
ID: 43766 · Report as offensive
Profile Eric J Korpela
Volunteer moderator
Project administrator
Project developer
Project scientist
Avatar

Send message
Joined: 15 Mar 05
Posts: 1547
Credit: 27,028,254
RAC: 1,959
United States
Message 43768 - Posted: 23 Sep 2012, 16:07:38 UTC - in response to Message 43766.  

Could you unhide your computers so I can see them without going directly to the database?
ID: 43768 · Report as offensive
Profile Sutaru Tsureku
Volunteer tester

Send message
Joined: 7 Jun 09
Posts: 285
Credit: 2,822,466
RAC: 0
Germany
Message 43769 - Posted: 23 Sep 2012, 16:42:49 UTC - in response to Message 43768.  
Last modified: 23 Sep 2012, 16:44:08 UTC

Done.

It look like it need some time before it work.

If I click in my above message to the host ID, my name is shown.

The host link in my account overview isn't shown until now.


- Best regards! :-) - Sutaru Tsureku, team seti.international founder. - Optimize your PC for higher RAC (@ SETI@home Main). - SETI@home needs your help. -
ID: 43769 · Report as offensive
Profile Sutaru Tsureku
Volunteer tester

Send message
Joined: 7 Jun 09
Posts: 285
Credit: 2,822,466
RAC: 0
Germany
Message 43770 - Posted: 23 Sep 2012, 17:02:28 UTC - in response to Message 43769.  
Last modified: 23 Sep 2012, 17:07:49 UTC

Sutaru Tsureku wrote:
(...)
The host link in my account overview isn't shown until now.


Also shown now.


- Best regards! :-) - Sutaru Tsureku, team seti.international founder. - Optimize your PC for higher RAC (@ SETI@home Main). - SETI@home needs your help. -
ID: 43770 · Report as offensive
Profile Eric J Korpela
Volunteer moderator
Project administrator
Project developer
Project scientist
Avatar

Send message
Joined: 15 Mar 05
Posts: 1547
Credit: 27,028,254
RAC: 1,959
United States
Message 43771 - Posted: 23 Sep 2012, 17:18:32 UTC - in response to Message 43770.  

I don't understand those "past deadline" issues at all. Our deadlines are 14 days, for SETI@home and 30 for Astropulse. Has the client or server been changed to impose different deadlines based on processing rate? If either is the case, that's a BOINC problem and an incredibly bad idea. We'll have to bring it to the BOINC projects list.
ID: 43771 · Report as offensive
Profile Sutaru Tsureku
Volunteer tester

Send message
Joined: 7 Jun 09
Posts: 285
Credit: 2,822,466
RAC: 0
Germany
Message 43772 - Posted: 23 Sep 2012, 17:50:41 UTC - in response to Message 43771.  

I don't know if the following is the problem ..

I made a few mistakes and my machine got a few new host ID's.

The WUs gone lost.

After I merged all hosts in the overview to only one host.

Maybe because of this, this all mixed up?


- Best regards! :-) - Sutaru Tsureku, team seti.international founder. - Optimize your PC for higher RAC (@ SETI@home Main). - SETI@home needs your help. -
ID: 43772 · Report as offensive
Profile Sutaru Tsureku
Volunteer tester

Send message
Joined: 7 Jun 09
Posts: 285
Credit: 2,822,466
RAC: 0
Germany
Message 43776 - Posted: 23 Sep 2012, 19:20:05 UTC - in response to Message 43772.  
Last modified: 23 Sep 2012, 19:27:10 UTC

Ahh .., maybe I know the problem with the very short deadline ..

Example:
http://setiweb.ssl.berkeley.edu/beta/result.php?resultid=10991838


postid=43198
Ohhh... they were GPU work units - it looks like...
To now I had only GPU enabled - and in past only this entries in app_info.xml file.

As BOINC asked only for GPU work units - it worked.

Then I enabled also CPU, then they all:
SETI@home Beta Test 06.07.2012 21:58:23 Didn't resend lost task xxxx (expired)


I guess it's the same like at SETI@home ..
BOINC ask for CPU WUs, the server would like to send .vlar WUs.
But the message from the server to BOINC go lost on the way.
BOINC don't download WUs.
The next time BOINC contact the server and ask for NVIDIA GPU WUs, the re-sent function is enabled, all the .vlar WUs get a very short deadline (*Didn't resend lost task xxxx (expired)*) - because .vlar WUs not to NVIDIA GPUs.


So I guess nothing wrong here ..


- Best regards! :-) - Sutaru Tsureku, team seti.international founder. - Optimize your PC for higher RAC (@ SETI@home Main). - SETI@home needs your help. -
ID: 43776 · Report as offensive
Wembley
Volunteer tester
Avatar

Send message
Joined: 13 Nov 09
Posts: 12
Credit: 135,674
RAC: 0
United States
Message 43777 - Posted: 23 Sep 2012, 19:49:34 UTC

I reset the project and enabled new tasks today. Got 2 AP's. Both errored out with "197 (0xc5) EXIT_TIME_LIMIT_EXCEEDED"

http://setiweb.ssl.berkeley.edu/beta/result.php?resultid=11178709
http://setiweb.ssl.berkeley.edu/beta/result.php?resultid=11178703

ID: 43777 · Report as offensive
Grumpy Swede
Volunteer tester
Avatar

Send message
Joined: 10 Mar 12
Posts: 1666
Credit: 13,037,228
RAC: 12,180
Sweden
Message 43788 - Posted: 24 Sep 2012, 15:30:51 UTC
Last modified: 24 Sep 2012, 15:55:34 UTC

Error error error, on Application details for host 57176, and 57179.

57179 certainly haven't produced for 6.04 "Consecutive valid tasks 13554" (my poor little ION wouldn't be able to do that many before the next ice age)

Application details for host 57179

and 57176 haven't done for 6.04 "Consecutive valid tasks 1438" (which it may be able to finish in a couple of years, but certainly not this soon)

Application details for host 57176

Hugely inflated. This happened sometime over the night (CET)

Added: It keeps counting up "Consecutive valid tasks" at an alarming rate, without me finishing any new tasks at all. Something seems stuck in a death loop here. If only I got 700 in credits for all those tasks too :-)

Edit: It's a runaway train. ION 6.04's now "Consecutive valid tasks 13755"
and NVIDIA GeForce 315M now "Consecutive valid tasks 1637".
ID: 43788 · Report as offensive
Grumpy Swede
Volunteer tester
Avatar

Send message
Joined: 10 Mar 12
Posts: 1666
Credit: 13,037,228
RAC: 12,180
Sweden
Message 43792 - Posted: 24 Sep 2012, 18:29:10 UTC
Last modified: 24 Sep 2012, 18:31:39 UTC

Well, the ION finally stopped counting "Consecutive valid tasks" at 14087, and the Samsung NVIDIA GeForce 315M, stopped counting "Consecutive valid tasks" at 1959.

Two ice ages and the ION will have reached the crazy value, and in 734 days, the Samsung will have made it to 1959 :-)

Interesting bug, whatever it was. At least I can now download an almost endless amount of 6.04's, since "Max tasks per day" followed the crazy count upward.

:-)
ID: 43792 · Report as offensive
Profile Eric J Korpela
Volunteer moderator
Project administrator
Project developer
Project scientist
Avatar

Send message
Joined: 15 Mar 05
Posts: 1547
Credit: 27,028,254
RAC: 1,959
United States
Message 43794 - Posted: 24 Sep 2012, 20:41:25 UTC - in response to Message 43792.  

Congratulations! You found a validator bug, and it's fix was my first check in to the new git repostory.

Things should preceed more normally now.
ID: 43794 · Report as offensive
Previous · 1 · 2 · 3 · 4 · 5 · Next

Message boards : News : Experiment for server operations check...


 
©2020 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.