Look at this host!

Message boards : Number crunching : Look at this host!
Message board moderation

To post messages, you must log in.

AuthorMessage
Cobbler

Send message
Joined: 2 Aug 01
Posts: 15
Credit: 3,450,756
RAC: 0
Hungary
Message 44165 - Posted: 8 Nov 2004, 6:58:59 UTC


http://setiweb.ssl.berkeley.edu/results.php?hostid=123036

It downloaded more than 300 WUs and returned them ALL with client error after about a minute of processing. Can it be a cheater-want-to-be?

Cobbler
ID: 44165 · Report as offensive
Profile Mike Special Project $75 donor
Volunteer tester
Avatar

Send message
Joined: 17 Feb 01
Posts: 34258
Credit: 79,922,639
RAC: 80
Germany
Message 44187 - Posted: 8 Nov 2004, 7:48:44 UTC

Hi

No, all the errors was by downloading.
I think this host has to reset the project or reatach.

greetz Mike



With each crime and every kindness we birth our future.
ID: 44187 · Report as offensive
Tony Martin

Send message
Joined: 5 Dec 99
Posts: 91
Credit: 69,723
RAC: 0
United States
Message 44190 - Posted: 8 Nov 2004, 7:53:55 UTC - in response to Message 44165.  


http://setiweb.ssl.berkeley.edu/show_host_detail.php?hostid=123036
http://setiweb.ssl.berkeley.edu/hosts_user.php?userid=30335

The 1st web site shows the faulty computer the 2nd web site shows all the computers for this person. There is something wrong with the host #123036 it has 315 wu's that all have downloading errors. Need to take this computer off line untill it gets fixed.
ID: 44190 · Report as offensive
Tony Martin

Send message
Joined: 5 Dec 99
Posts: 91
Credit: 69,723
RAC: 0
United States
Message 44507 - Posted: 9 Nov 2004, 0:56:52 UTC

here is another host to look at
http://setiweb.ssl.berkeley.edu/show_host_detail.php?hostid=323127

this user has 280 results of which 249 haven't been run and will be in an over no reply state soon. just another person that has their connect to server settings set to high. a connect to server every 0.5 days is plenty of WU's to have on one computer. Seti hasn't had any server outages for some time now and there is no reason to hoard WU's.
ID: 44507 · Report as offensive
Scavenger

Send message
Joined: 10 Jun 99
Posts: 5
Credit: 272,422
RAC: 0
United States
Message 44510 - Posted: 9 Nov 2004, 1:10:10 UTC
Last modified: 4 Dec 2004, 3:15:25 UTC


ID: 44510 · Report as offensive
ChinookFoehn

Send message
Joined: 18 Apr 02
Posts: 462
Credit: 24,039
RAC: 0
Message 44514 - Posted: 9 Nov 2004, 1:16:15 UTC - in response to Message 44507.  
Last modified: 17 Dec 2004, 10:17:46 UTC

ID: 44514 · Report as offensive
Tony Martin

Send message
Joined: 5 Dec 99
Posts: 91
Credit: 69,723
RAC: 0
United States
Message 44520 - Posted: 9 Nov 2004, 1:22:58 UTC - in response to Message 44510.  

> How about this one. Shows to have 16,072 computers. All I checked had
> downloaded at least 1 work unit.
>
> http://setiweb.ssl.berkeley.edu/hosts_user.php?userid=701952
>
This user seems to have attached and detached 16069 times. All of the computers after his 3rd one seem to be the same computer so he must have attached and detached to get that many computers to show up for his account.
ID: 44520 · Report as offensive
Tony Martin

Send message
Joined: 5 Dec 99
Posts: 91
Credit: 69,723
RAC: 0
United States
Message 44522 - Posted: 9 Nov 2004, 1:25:53 UTC - in response to Message 44514.  

> > ... connect to server every 0.5 days is plenty of WU's to
> > have on one computer. Seti hasn't had any server outages for some time
> now and
> > there is no reason to hoard WU's.
>
> If you are using any processors with Hyper-Threading; then this setting is too
> low

It shouldn't matter if you have HT machines or not as soon as the machine gets to the point that you will run out of work in 0.5 days it will download more WU's.
ID: 44522 · Report as offensive
Profile Benher
Volunteer developer
Volunteer tester

Send message
Joined: 25 Jul 99
Posts: 517
Credit: 465,152
RAC: 0
United States
Message 44526 - Posted: 9 Nov 2004, 1:30:13 UTC

I found this in the error message section of one of the results (from original post)
"...setiathome_4.07_windows_intelx86.exe: signature verification error"

Can't recall ATM if that is the WU signature or the application itself.
I know you can compile your own version of the application, but if you just put it into the projects folder the boinc application will reject it (wrong signature).
ID: 44526 · Report as offensive
ChinookFoehn

Send message
Joined: 18 Apr 02
Posts: 462
Credit: 24,039
RAC: 0
Message 44527 - Posted: 9 Nov 2004, 1:32:23 UTC - in response to Message 44522.  
Last modified: 17 Dec 2004, 10:18:20 UTC

ID: 44527 · Report as offensive
ChinookFoehn

Send message
Joined: 18 Apr 02
Posts: 462
Credit: 24,039
RAC: 0
Message 44532 - Posted: 9 Nov 2004, 1:42:38 UTC - in response to Message 44510.  
Last modified: 17 Dec 2004, 10:18:03 UTC

ID: 44532 · Report as offensive
1mp0£173
Volunteer tester

Send message
Joined: 3 Apr 99
Posts: 8423
Credit: 356,897
RAC: 0
United States
Message 44535 - Posted: 9 Nov 2004, 1:49:07 UTC - in response to Message 44527.  

2 days seems like a really good number. BOINC will keep 2 to 4 days of work, and if things are down, well, they're down.

Numbers bigger than about 5 days should probably not be allowed.

> Not if, as was my case, LHC and Seti were down for several days and CPDN had
> their first occurrance of the w.u. server locking up all week-end.
>
> LHC still shuts down every night from midnight C.E.T. until about 06:00 C.E.T.
> (23:00-05:00 Zulu) while no-one can convince me that Seti will never crash
> again, while CPDN still, occasionally, locks up - usually on weekends. Until 5
> projects are up and running, I do not advise a lower for H-T enabled
> processors.
>
ID: 44535 · Report as offensive
Scavenger

Send message
Joined: 10 Jun 99
Posts: 5
Credit: 272,422
RAC: 0
United States
Message 44537 - Posted: 9 Nov 2004, 1:54:23 UTC - in response to Message 44532.  
Last modified: 4 Dec 2004, 3:14:49 UTC


ID: 44537 · Report as offensive
ChinookFoehn

Send message
Joined: 18 Apr 02
Posts: 462
Credit: 24,039
RAC: 0
Message 44546 - Posted: 9 Nov 2004, 2:12:05 UTC - in response to Message 44535.  
Last modified: 17 Dec 2004, 10:17:31 UTC

ID: 44546 · Report as offensive
Ingleside
Volunteer developer

Send message
Joined: 4 Feb 03
Posts: 1546
Credit: 15,832,022
RAC: 13
Norway
Message 44557 - Posted: 9 Nov 2004, 2:38:59 UTC - in response to Message 44546.  

>
> Perhaps I am wrong, but I thought if one is attached to say 3 projects, and
> has the setting at 1 day, then 1 day's of work would be downloaded from each
> project thereby providing 3 days worth of work - assuming all projects are set
> the same resource share.

No, you'll download 1 day with current share from each project, meaning roughly 8 hours cpu-time from each if 33% share in 3 projects.
If example seti-wu is expected to take 2 hours and lhc 5 hours you'll most likely download 5 seti & 2 lhc & 1 cpdn.

Well, actually you'll download roughly 2 days of work from each project, since the "connect every n days" gives you roughly 2n days cached, this means 9 seti & 4 lhc & 1 cpdn.

>
> Of course, if one is attached to CPDN, this becomes moot until towards the end
> of the CPDN work unit.
>

If you're not so unlucky CPDN crashes for some reason.
ID: 44557 · Report as offensive
ChinookFoehn

Send message
Joined: 18 Apr 02
Posts: 462
Credit: 24,039
RAC: 0
Message 44569 - Posted: 9 Nov 2004, 3:00:07 UTC - in response to Message 44557.  
Last modified: 17 Dec 2004, 10:17:15 UTC

ID: 44569 · Report as offensive
Profile Jord
Volunteer tester
Avatar

Send message
Joined: 9 Jun 99
Posts: 15184
Credit: 4,362,181
RAC: 3
Netherlands
Message 44688 - Posted: 9 Nov 2004, 9:58:03 UTC - in response to Message 44557.  

> > Of course, if one is attached to CPDN, this becomes moot until towards
> the end
> > of the CPDN work unit.
> >
>
> If you're not so unlucky CPDN crashes for some reason.
>
What? I am not the only one whose CPDN units started crashing (with -5 errors) as soon as the Beta unit (crunched with CPDN 4.02) was finished and the CPDN cruncher updated to 4.04?

Carl already told me that I may have to test one or two more CPDN units. If they keep crashing, my only hope would be another BOINC version. :(

Do watch out if CPDN crashes, for the hadsm3um will continue to run in the background.
ID: 44688 · Report as offensive
Ingleside
Volunteer developer

Send message
Joined: 4 Feb 03
Posts: 1546
Credit: 15,832,022
RAC: 13
Norway
Message 44751 - Posted: 9 Nov 2004, 14:17:55 UTC - in response to Message 44688.  

>
> What? I am not the only one whose CPDN units started crashing (with -5 errors)
> as soon as the Beta unit (crunched with CPDN 4.02) was finished and the CPDN
> cruncher updated to 4.04?
>
> Carl already told me that I may have to test one or two more CPDN units. If
> they keep crashing, my only hope would be another BOINC version. :(
>
> Do watch out if CPDN crashes, for the hadsm3um will continue to run in the
> background.
>

Going a little off-topic here, but have also had a couple CPDN-crashes. But one crash was own stupidity, so can't really blame BOINC or CPDN for this. The 2nd was another program having a memory-leak, and during the night ran out of all memory & pagefile so both CPDN-wu crashed. :(

Is currently crunching both v4.03 & v4.04, and haven't had any problems.
ID: 44751 · Report as offensive

Message boards : Number crunching : Look at this host!


 
©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.