Continued server problems.


log in

Advanced search

Message boards : News : Continued server problems.

Previous · 1 · 2 · 3
Author Message
Profile Jimbo
Volunteer tester
Avatar
Send message
Joined: 21 Jun 99
Posts: 1
Credit: 924,680
RAC: 648
United States
Message 1308679 - Posted: 22 Nov 2012, 8:15:49 UTC
Last modified: 22 Nov 2012, 8:20:34 UTC

'AS' I was downloading work units (from seti@home) my antivirus window popped up and quarantined the following !!
i hope there are no serious bugs..,

Risk name: Trojan.Win32.Encpk.da(v) Location: ProgramData\Boinc\projects\setiathome.berkeley.edu\libfftw3f-3-1-1a_upx.dll
____________

Profile dancer42
Volunteer tester
Send message
Joined: 2 Jun 02
Posts: 436
Credit: 1,160,474
RAC: 80
United States
Message 1308684 - Posted: 22 Nov 2012, 8:49:15 UTC - in response to Message 1308603.

First to the staff i realize that what you are doing is hard so thanks for the hard work and good luck.
Second would it be possible to move the comment boards to a different server so when things crash on the work side we can still find out what to do or what is going on wright now in bad crashes the boards go down too.
Third 1 or 2 hour jobs for what, my hd6870 will run 24 astropulse jobs a day my5770 12 and my 5770's 4 jobs a day each my 8 core cpu will run about 1 a day per core.
you can in preferences set the number of core's and gpu's, yes or no to using cpu,gpu the % of cpu and gpu time that can be used so if running at 100% makes work difficult then a adjust the preferences.
There are many of us out here that have saved our nickels and dimes and have stuffed a better gpu or two or three or have built second machines so the larger work unit is welcome.
If you are adamant about the work unit size then help support the raising of enough money to endow 1 full time programmer that could take care of things like a vary able size work unit based on benchmark or just donate because it is a good thing!
____________

Profile dancer42
Volunteer tester
Send message
Joined: 2 Jun 02
Posts: 436
Credit: 1,160,474
RAC: 80
United States
Message 1308690 - Posted: 22 Nov 2012, 8:59:52 UTC - in response to Message 1308679.

looks like you may be infected
It is unlikely the seti@home server infected you it is running unix os,but theres is a vary vary small chance of cross platform infection any body else see this?
After looking this up you probably were infected by a fake flash file it looks nasty if you can wipe system with out losing important file's recomend wipe and reload.
____________

juan BFBProject donor
Volunteer tester
Avatar
Send message
Joined: 16 Mar 07
Posts: 5471
Credit: 313,433,666
RAC: 118,577
Brazil
Message 1308695 - Posted: 22 Nov 2012, 9:17:56 UTC
Last modified: 22 Nov 2012, 9:22:15 UTC

On this host, with out proxy still canĀ“t DL almost nothing (DL <1kbps), with proxy (10-20kbps), Scheduler Failure 31%, Scheduler Timeouts 3% of total and 12% of failures.

Are the kitties able to DL anything directly? E@H DL directly at >1Mbps so i belive my ISP was doing its job well.
____________

Profile dancer42
Volunteer tester
Send message
Joined: 2 Jun 02
Posts: 436
Credit: 1,160,474
RAC: 80
United States
Message 1308696 - Posted: 22 Nov 2012, 9:18:42 UTC - in response to Message 1308509.

Basically each task seti sends out occupies one spot in a master data base.
and for any number of reasons system crashes internet problems and so on some of these tasks ether can not be done or are abandon.
this number has gotten so large and the systems have gotten so fast that the scheduling server can not keep up.
Seti has limited the jobs to 100 at a time so that most of the dead files can be purged and rescheduled,hence weeks to mouths to fix, 3 weeks for first timeout pass less time for 2 and three.
____________

Profile S@NL Blue Angel
Avatar
Send message
Joined: 11 May 03
Posts: 224
Credit: 4,545,031
RAC: 19
Netherlands
Message 1308720 - Posted: 22 Nov 2012, 10:26:57 UTC
Last modified: 22 Nov 2012, 10:27:31 UTC

Just to give everyone who don't understand a tip:
When you don't get work shut down the button of I want work
for a while, then push on project and wait till the done WU's
are gone, and then click on the I want work botton again
you will see the task will run. At my computer when I have
20 task/wu done I'll get 20 task back for download.
Atleast the program is keep on running;)

Greetings Blue_Angel
____________

juan BFBProject donor
Volunteer tester
Avatar
Send message
Joined: 16 Mar 07
Posts: 5471
Credit: 313,433,666
RAC: 118,577
Brazil
Message 1308725 - Posted: 22 Nov 2012, 10:41:46 UTC - in response to Message 1308720.

Just to give everyone who don't understand a tip:
When you don't get work shut down the button of I want work
for a while, then push on project and wait till the done WU's
are gone, and then click on the I want work botton again
you will see the task will run. At my computer when I have
20 task/wu done I'll get 20 task back for download.
Atleast the program is keep on running;)

Greetings Blue_Angel

My problem is thats kind of tip does not works on hungry big rings, unless you do a lot of babysitting, they need to be feeded allmost all time due the 100WU limit.

Because thats i need a stable DL condition, free proxy works but i use to much bandwidth, and i allways be kicked out after some time, that leaves to more babysitting.

I agree that is not a common situation, your tip works fine if you have few hosts that needs to be feeded with few WU per day, is a good "band-aid" solution for most of the users.
____________

Richard HaselgroveProject donor
Volunteer tester
Send message
Joined: 4 Jul 99
Posts: 8760
Credit: 52,712,913
RAC: 21,630
United Kingdom
Message 1308728 - Posted: 22 Nov 2012, 11:09:57 UTC

Guys, guys.....

Please stop rehashing old (and by now very outdated) problems in the News area.

The scheduler problems were solved - completely, as far as I can tell from here - over 12 hours ago.

Please check any virus warnings with a service like virustotal.com. If you find a real virus, that would indeed be important news - but so far every warning has turned out to be a false alarm.

Any remaining issues/questions are far better aired in either Number crunching or Questions and answers.

Thank you.

Profile S@NL Blue Angel
Avatar
Send message
Joined: 11 May 03
Posts: 224
Credit: 4,545,031
RAC: 19
Netherlands
Message 1308729 - Posted: 22 Nov 2012, 11:12:42 UTC - in response to Message 1308725.
Last modified: 22 Nov 2012, 11:14:05 UTC

Just to give everyone who don't understand a tip:
When you don't get work shut down the button of I want work
for a while, then push on project and wait till the done WU's
are gone, and then click on the I want work botton again
you will see the task will run. At my computer when I have
20 task/wu done I'll get 20 task back for download.
Atleast the program is keep on running;)

Greetings Blue_Angel

My problem is thats kind of tip does not works on hungry big rings, unless you do a lot of babysitting, they need to be feeded allmost all time due the 100WU limit.

Because thats i need a stable DL condition, free proxy works but i use to much bandwidth, and i allways be kicked out after some time, that leaves to more babysitting.

I agree that is not a common situation, your tip works fine if you have few hosts that needs to be feeded with few WU per day, is a good "band-aid" solution for most of the users.


I have bandwidth to, I have a windows 5i close to 7
Maybe you understood me wrong but my task keep staying
on 100, never less. It fill every time a wu is done
the download run again.
____________

Profile S@NL Blue Angel
Avatar
Send message
Joined: 11 May 03
Posts: 224
Credit: 4,545,031
RAC: 19
Netherlands
Message 1308730 - Posted: 22 Nov 2012, 11:17:10 UTC - in response to Message 1308728.
Last modified: 22 Nov 2012, 11:17:25 UTC

Guys, guys.....

Please stop rehashing old (and by now very outdated) problems in the News area.

The scheduler problems were solved - completely, as far as I can tell from here - over 12 hours ago.

Please check any virus warnings with a service like virustotal.com. If you find a real virus, that would indeed be important news - but so far every warning has turned out to be a false alarm.

Any remaining issues/questions are far better aired in either Number crunching or Questions and answers.

Thank you.


Richard you are right,

I'll wrote this for people who don't understand.
But Have a nice Tanks Giving all!!
____________

Profile Ageless
Avatar
Send message
Joined: 9 Jun 99
Posts: 12396
Credit: 2,666,376
RAC: 789
Netherlands
Message 1308737 - Posted: 22 Nov 2012, 12:06:20 UTC - in response to Message 1308730.

Tanks Giving

.... Now there's a typo.


____________
Jord

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

Profile Donald L. JohnsonProject donor
Avatar
Send message
Joined: 5 Aug 02
Posts: 6326
Credit: 769,184
RAC: 749
United States
Message 1308794 - Posted: 22 Nov 2012, 15:14:29 UTC - in response to Message 1308684.
Last modified: 22 Nov 2012, 15:20:08 UTC

First to the staff i realize that what you are doing is hard so thanks for the hard work and good luck.
Second would it be possible to move the comment boards to a different server so when things crash on the work side we can still find out what to do or what is going on wright now in bad crashes the boards go down too.

If I understand the system correctly, the Forums actually ARE on a separate server - thinman (data-driven web pages). But since posting to the Forums requires a minimum S@H RAC, and Forum messages show some of the poster's S@H stats, every Forum access (read or post) requires access to the BOINC database. That access adds to the load on the database server (DB Queries/second on the Server Stats page). I think the individual messages may also be stored in the database, but in a different section from the Work Unit tracking.
____________
Donald
Infernal Optimist / Submariner, retired

Omega Man
Send message
Joined: 28 Dec 99
Posts: 2
Credit: 1,737,204
RAC: 1,646
United States
Message 1308836 - Posted: 22 Nov 2012, 17:26:46 UTC - in response to Message 1307567.

I updated BOINC 3 days ago and still have not received a single WU.

Eric KorpelaProject donor
Volunteer moderator
Project administrator
Project developer
Project scientist
Avatar
Send message
Joined: 3 Apr 99
Posts: 1111
Credit: 10,323,780
RAC: 10,907
United States
Message 1308837 - Posted: 22 Nov 2012, 17:27:24 UTC - in response to Message 1308085.

Good idea. Thanks!
____________

Profile dancer42
Volunteer tester
Send message
Joined: 2 Jun 02
Posts: 436
Credit: 1,160,474
RAC: 80
United States
Message 1309004 - Posted: 22 Nov 2012, 21:25:52 UTC

scheduling server is down let eric enjoy turkey day we can bug him tomorrow.
____________

Previous · 1 · 2 · 3

Message boards : News : Continued server problems.

Copyright © 2014 University of California