Problem reporting tasks

Message boards : Number crunching : Problem reporting tasks
Message board moderation

To post messages, you must log in.

Previous · 1 . . . 4 · 5 · 6 · 7

AuthorMessage
Swamirabbitina
Volunteer tester

Send message
Joined: 16 Sep 05
Posts: 2
Credit: 409,583
RAC: 0
United States
Message 1021825 - Posted: 2 Aug 2010, 0:53:24 UTC

same problem with me. Getting lots of Scheduler request failed: Http internal server error and also the Can't connect message, I have 42 WU backed up. funny thing is that WU upload just fine; just not getting credit, which hasn't changed for several days.

where is this app_info.xml file supposed to be located? I can't find it. running 6.11.4 32 bit on windows 7/
ID: 1021825 · Report as offensive
Profile perryjay
Volunteer tester
Avatar

Send message
Joined: 20 Aug 02
Posts: 3377
Credit: 20,676,751
RAC: 0
United States
Message 1021826 - Posted: 2 Aug 2010, 1:04:07 UTC - in response to Message 1021825.  

Hi Swami, welcome to the forums,

You wont find an app_info, you are not running optimized apps. You might want to run over to the Lunatic's site http://lunatics.kwsn.net/index.php?module=Downloads;catd=9 and see about trying it though. Optimized apps can cut your crunch time dramatically. You may need to run CPUZ also just to make sure what installer to use though I see you are running AMD so your choices may be limited.


PROUD MEMBER OF Team Starfire World BOINC
ID: 1021826 · Report as offensive
Profile skildude
Avatar

Send message
Joined: 4 Oct 00
Posts: 9541
Credit: 50,759,529
RAC: 60
Yemen
Message 1021838 - Posted: 2 Aug 2010, 3:08:36 UTC - in response to Message 1021826.  
Last modified: 2 Aug 2010, 3:36:37 UTC

heres something interesting I saw on my Mandriva box.

I installed the OS upgrade and needed to tweak BOINC a bit. well long story short BOINC the OS somehow lost the apps and App_info.xml that I installed. so after much frustration about communication errors I just detached from seti and reattached. No harm since my WU's failed from the already mentioned missing files. Turns out it got work. So far so good. I can connect to seti and get work. I then stopped the BOINC process installed the AK MB and beta4 AP apps to seti@home and then restarted boinc. WU dies. then I get the HTTP error. THis is definitely something server side that is quite annoying.

On a happy note, Primegrid is getting a lot of work tossed their way until this clears

/EDIT
I got work on the Mandriva box from seti. So, apparently its just certain boxes that cant get through


In a rich man's house there is no place to spit but his face.
Diogenes Of Sinope
ID: 1021838 · Report as offensive
Profile [B^S] madmac
Volunteer tester
Avatar

Send message
Joined: 9 Feb 04
Posts: 1175
Credit: 4,754,897
RAC: 0
United Kingdom
Message 1021852 - Posted: 2 Aug 2010, 6:35:26 UTC
Last modified: 2 Aug 2010, 6:42:11 UTC

I have got two machines one running Windows7 and the other XP it is mainly the XP one I get internal server errors. The Windows 7 sometimes I get the message or I get cannot connect to server internet access ok so it must be their end the problem is. Both are uning the Lunatics App file only had this problem since the update, cannot get a latest score for my XP machine all I get is internal server errors.How do I get rid of the v5.00 version from my app file, file only update 13th July
ID: 1021852 · Report as offensive
Profile [B^S] madmac
Volunteer tester
Avatar

Send message
Joined: 9 Feb 04
Posts: 1175
Credit: 4,754,897
RAC: 0
United Kingdom
Message 1021855 - Posted: 2 Aug 2010, 8:01:19 UTC - in response to Message 1021852.  

I have got two machines one running Windows7 and the other XP it is mainly the XP one I get internal server errors. The Windows 7 sometimes I get the message or I get cannot connect to server internet access ok so it must be their end the problem is. Both are uning the Lunatics App file only had this problem since the update, cannot get a latest score for my XP machine all I get is internal server errors.How do I get rid of the v5.00 version from my app file, file only update 13th July

Further info thought I had different versions yet both are running 10.56 should I update and see if this clears it
ID: 1021855 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14649
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1021856 - Posted: 2 Aug 2010, 8:14:26 UTC - in response to Message 1021659.  

I used the Lunatics site, should I change my info file for the Astropulse v5.00?

Doesn't really matter what you used: there have been so many changes, so many different versions, so many suggestions for things that that you might or might not change yourself, that the only acid test is for you to look inside your own copy and see what you find there.

IF you're getting this specific error message ("HTTP Internal Server Error") when you try to report or request new work, then yes, you've quite possibly got a old application name in there, and removing it will help resolve the problem more quickly.

But hopefully my bug report will alert someone: David Anderson sometimes codes on a Sunday evening, and it should be easy to fix, tomorrow if not today. You could just wait for that to happen.

I see changeset [trac]changeset:22090[/trac]:

Timestamp: 08/01/10 20:43:38 (4 hours ago)
Author: davea
Message: scheduler: fix bug that crashed if app_info.xml refers to nonexistent app

If he was coding at 8 o'clock on a Sunday evening (timesramp is local time, not UTC), it's unlikely it'll be live on the server yet. But hopefully they'll be able to fix it quickly once the lab opens.

ID: 1021856 · Report as offensive
Lionel

Send message
Joined: 25 Mar 00
Posts: 680
Credit: 563,640,304
RAC: 597
Australia
Message 1021857 - Posted: 2 Aug 2010, 8:41:42 UTC - in response to Message 1021602.  

I found that in my case the HTTP Server internal error was related to my app_info.xml file. This contained references to older application versions which are no longer used. (the content of app_info.xml is reported on any update)
It seems to me that the server update from the 26th of July caused some obsolete application versions to be invalid. In stead of reporting an invalid application it crashes causing a internal server error.

Removing the references to the obsolete applications from my app_info.xml solved this issue for me. The next update completed succesfully and my wu's are reported.


removing the old entries in the app info file works ... thnx very much ... the lappy is happy again.

L.
ID: 1021857 · Report as offensive
Profile Bill Walker
Avatar

Send message
Joined: 4 Sep 99
Posts: 3868
Credit: 2,697,267
RAC: 0
Canada
Message 1021883 - Posted: 2 Aug 2010, 13:10:41 UTC

There may be two seperate issues. Removing old ap references on ap.info got ride of the "http error", but I was still getting "cannot connect to server". Deatached, reattached and got one WU, estimating 25 hours to complete (forgot to reinstall optimized aps). After using the Lunatics installer, got 27 WUs, estimating 0:40 to 2:30 to complete (which is pretty close to actual). I'm almost set for the next outrage.

ID: 1021883 · Report as offensive
Robert Ribbeck
Avatar

Send message
Joined: 7 Jun 02
Posts: 644
Credit: 5,283,174
RAC: 0
United States
Message 1021889 - Posted: 2 Aug 2010, 13:32:16 UTC - in response to Message 1021883.  

There may be two seperate issues. Removing old ap references on ap.info got ride of the "http error", but I was still getting "cannot connect to server". Deatached, reattached and got one WU, estimating 25 hours to complete (forgot to reinstall optimized aps). After using the Lunatics installer, got 27 WUs, estimating 0:40 to 2:30 to complete (which is pretty close to actual). I'm almost set for the next outrage.


I'd have to say it's a separate issue
When I checked my ap.info it was correct but I've had trouble with connections reporting. I think it's something they did last outage. Reporting worked fine be for the last outage
ID: 1021889 · Report as offensive
Profile skildude
Avatar

Send message
Joined: 4 Oct 00
Posts: 9541
Credit: 50,759,529
RAC: 60
Yemen
Message 1021894 - Posted: 2 Aug 2010, 13:42:23 UTC - in response to Message 1021889.  

Agreed I don't have old application data in my App_info so this seems to be a red herring


In a rich man's house there is no place to spit but his face.
Diogenes Of Sinope
ID: 1021894 · Report as offensive
Josef W. Segur
Volunteer developer
Volunteer tester

Send message
Joined: 30 Oct 99
Posts: 4504
Credit: 1,414,761
RAC: 0
United States
Message 1021910 - Posted: 2 Aug 2010, 14:13:22 UTC - in response to Message 1021764.  

With the "internal server error" issue understood and being fixed, here's some thoughts on the scheduler request failed: Couldn't connect to server problem.

Hellsheep wrote:
Hey guys and girls,

Sorry i haven't been around much. Been busy working on some web servers for a company.

Anyway.. After having these same issues i did some debugging.

Noticed that the URL boinc tries to connect to is initiating a request on port 80 (http) so naturally that means apache or something is installed on that server.

Tried accessing http://setiboinc.ssl.berkeley.edu or even http://ssl.berkeley.edu a few times over an hour, i got 4 successful requests which returned the default fedora test page. And the rest were not successful.

"The server at http://setiboinc.ssl.berkeley.edu is taking too long to respond."
...

The question, in part, is what "too long" means.

A TCP connection attempt involves sending a SYN packet and waiting for the ACK. Windows implementation waits 3 seconds for the ACK, then tries another SYN, waits 6 seconds, etc. doubling the allowed delay each time. Windows NT and Win9x default to allowing 3 retries so the final failure takes a total of 3+6+12+24 = 45 seconds. Windows 2000 and later default to 2 retries, so it's 21 seconds.

From my eastern US location, I've been getting some "couldn't connect" failures, though not enough to be a serious problem. Using a web browser on my Win98SE daily driver to get http://setiboinc.ssl.berkeley.edu a dozen times over about an hour's time was always successful, but took from 3 seconds to 1 minute 36 seconds. The web browser tries again after the first failure to connect, about half those trials took longer than the Win98SE 45 second connection timeout.

The Windows registry value which controls how many SYN retries are done is TcpMaxConnectRetransmissions. If I were having serious "couldn't connect" problems I'd be considering increasing it to 4 which would give a 93 second timeout.

There was a "21 second timeout" thread back in 2007 related to connecting to the download server, increasing TcpMaxConnectRetransmissions in that case didn't seem to help. Possibly later versions of Windows ignore it, or possibly whatever was interfering with connections lasted longer than any reasonable timeout sequence.
                                                                   Joe
ID: 1021910 · Report as offensive
Profile Searcher
Avatar

Send message
Joined: 26 Jun 99
Posts: 139
Credit: 9,063,168
RAC: 15
United States
Message 1021923 - Posted: 2 Aug 2010, 14:57:01 UTC
Last modified: 2 Aug 2010, 14:59:40 UTC

I have had the same issue as the OP since about the 25th. After reading thru this thread I upgraded to the latest enhanced app and all 80 reports uploaded. The time stamp on the app_info.xml did not change.
ID: 1021923 · Report as offensive
Profile skildude
Avatar

Send message
Joined: 4 Oct 00
Posts: 9541
Credit: 50,759,529
RAC: 60
Yemen
Message 1021941 - Posted: 2 Aug 2010, 15:59:29 UTC - in response to Message 1021923.  

Hmmm thats interesting. reinstalled my opt apps using lunatics unified installer 0.36 and all my work reported and I got tasks downloading. That seems like a silly fix but it does work


In a rich man's house there is no place to spit but his face.
Diogenes Of Sinope
ID: 1021941 · Report as offensive
Rob.B

Send message
Joined: 23 Jul 99
Posts: 157
Credit: 1,439,682
RAC: 0
United Kingdom
Message 1021950 - Posted: 2 Aug 2010, 16:14:58 UTC

I have done most of what has been said perviously and i still get;

02/08/2010 17:10:56 SETI@home Reporting 11 completed tasks, not requesting new tasks
02/08/2010 17:11:18 Project communication failed: attempting access to reference site
02/08/2010 17:11:19 Internet access OK - project servers may be temporarily down.
02/08/2010 17:11:21 SETI@home Scheduler request failed: Failure when receiving data from the peer

So the question is what has changed? I know it is not me, so it's else where. If so whrere is the prior notification, clock cycles ar enot free, some one has to pay the electricity bill. I know I'll get a kicking from the sycophants, but it has to be said.

Rob.B
ID: 1021950 · Report as offensive
Profile skildude
Avatar

Send message
Joined: 4 Oct 00
Posts: 9541
Credit: 50,759,529
RAC: 60
Yemen
Message 1021959 - Posted: 2 Aug 2010, 16:37:50 UTC - in response to Message 1021950.  

reinstalling with the newer 0.36 lunatics unified installer didnt help? did you turn off boinc before the install?


In a rich man's house there is no place to spit but his face.
Diogenes Of Sinope
ID: 1021959 · Report as offensive
Profile [B^S] madmac
Volunteer tester
Avatar

Send message
Joined: 9 Feb 04
Posts: 1175
Credit: 4,754,897
RAC: 0
United Kingdom
Message 1021987 - Posted: 2 Aug 2010, 18:09:13 UTC

I reinstalled the new cersion of Lunatics 0.36 hit update to see what happens and I got everything ok it updated my score whilst the previous version I got internal server errors etc, hopefully everything is ok when I restarted Seti when my climate runs out on my oldest machine
ID: 1021987 · Report as offensive
Profile [B^S] madmac
Volunteer tester
Avatar

Send message
Joined: 9 Feb 04
Posts: 1175
Credit: 4,754,897
RAC: 0
United Kingdom
Message 1022041 - Posted: 2 Aug 2010, 21:11:06 UTC
Last modified: 2 Aug 2010, 21:11:36 UTC

either I am getting no new work messages or cannot connect to server messages, will just sit and hopefully the scheduler problem will be sorted out. Getting these messages from 18:00 BST Forgot to add this is on my Windows 7 machine
ID: 1022041 · Report as offensive
Previous · 1 . . . 4 · 5 · 6 · 7

Message boards : Number crunching : Problem reporting tasks


 
©2024 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.