Project: 3051 (Run 6, Clone 24, Gen 51)

Moderators: Site Moderators, FAHC Science Team

Post Reply
Newbie1Kenobi
Posts: 8
Joined: Thu Dec 06, 2007 3:14 am

Project: 3051 (Run 6, Clone 24, Gen 51)

Post by Newbie1Kenobi »

My quad is going at this WU for the sixth consecutive time. Every other time it has always EUE'd at 23%. Is it safe to assume this is a bad WU?

Specs are:

Q6600-not OC'd
2 gb DDRII 533 RAM
Ubuntu 7.04
Ivoshiee
Site Moderator
Posts: 822
Joined: Sun Dec 02, 2007 12:05 am
Location: Estonia

Post by Ivoshiee »

An EUE at the same point can be an indicator for the bad WU, but is it issue of the WU itself or only your computer, can only be confirmed by running it on some other computer.
gwildperson
Posts: 450
Joined: Tue Dec 04, 2007 8:36 pm

Post by gwildperson »

If I see the same error at the same point after several tries at the same WU, I delete the WU and move on. That's what the message says, after all: Deleting WU and moving on.
Ivoshiee
Site Moderator
Posts: 822
Joined: Sun Dec 02, 2007 12:05 am
Location: Estonia

Post by Ivoshiee »

gwildperson wrote:If I see the same error at the same point after several tries at the same WU, I delete the WU and move on. That's what the message says, after all: Deleting WU and moving on.
If the WU is crashing at the same point with the message about "UNKNOWN ERROR" then it will be good to have this WU tested on some other system. If it does crash there too then it may be possible to research the cause of it and turn it into a regular EUE event.
For example:
http://foldingforum.org/viewtopic.php?t=258
Newbie1Kenobi
Posts: 8
Joined: Thu Dec 06, 2007 3:14 am

Post by Newbie1Kenobi »

Sorry..I should have made a backup of my files. I didn't :oops: I deleted my work folder and queue file and got a new WU...I've gone through two or three with no errors now, so I have to assume it was a bad WU.

If it happens again, I'll make sure to have a backup made, in case someone else wants to try running it.
_ikki_
Posts: 27
Joined: Wed Dec 05, 2007 8:38 am

Post by _ikki_ »

Newbie1Kenobi wrote:Sorry..I should have made a backup of my files. I didn't :oops: I deleted my work folder and queue file and got a new WU...I've gone through two or three with no errors now, so I have to assume it was a bad WU.

If it happens again, I'll make sure to have a backup made, in case someone else wants to try running it.
_ikki_ wrote:The next time I will be ready :)

Code: Select all

#!/bin/bash
# /home/ikki/fah/backup.sh
# script to backup FAH files

suffix=$(date +%s)
cd /home/ikki/fah && tar cvzf "fah-$suffix.tgz" inst1/ >/dev/null

Code: Select all

#cron task
0 * * * * /home/ikki/fah/backup.sh
It will be useful :)
Ivoshiee
Site Moderator
Posts: 822
Joined: Sun Dec 02, 2007 12:05 am
Location: Estonia

Post by Ivoshiee »

_ikki_ wrote:
Newbie1Kenobi wrote:Sorry..I should have made a backup of my files. I didn't :oops: I deleted my work folder and queue file and got a new WU...I've gone through two or three with no errors now, so I have to assume it was a bad WU.

If it happens again, I'll make sure to have a backup made, in case someone else wants to try running it.
_ikki_ wrote:The next time I will be ready :)

Code: Select all

#!/bin/bash
# /home/ikki/fah/backup.sh
# script to backup FAH files

suffix=$(date +%s)
cd /home/ikki/fah && tar cvzf "fah-$suffix.tgz" inst1/ >/dev/null

Code: Select all

#cron task
0 * * * * /home/ikki/fah/backup.sh
It will be useful :)
The FAH *nix installation script called the finstall will bundle the FAH client backup script with itself. You'll get something similar:
#!/bin/bash
#
# ./fahback is for backing up FAH installation.
# To get to know more about FAH go to http://folding.stanford.edu.
#
# Primary reason of this is to help catch "BADWU1" type of WUs.
# "BADWU1" type WUs are these Gromacs WUs what will freeze/lockup the AMD CPUs when SSE is enabled.
# For more information:
# http://forum.folding-community.org/viewtopic.php?t=4444
#
# To get periodic backups (30min):
# Run "crontab -e", type "i" and add this line:
# */30 * * * * /home/ivo/public_html/fahback
# Then press [esc] and ":x".
#

Fpath="/home/ivo/foldingathome"
FAH="/home/ivo/public_html"

if [ -r ${FAH}/fah_back1.tgz ];
then
if [ -r ${FAH}/fah_back2.tgz ];
then
rm -f ${FAH}/fah_back2.tgz
fi
mv ${FAH}/fah_back1.tgz ${FAH}/fah_back2.tgz
fi
tar czf ${FAH}/fah_back1.tgz ${Fpath}
More information about the finstall script can be found all over the WWW, but for starters:
http://ra.vendomar.ee/%7Eivo/finstall_in_action.html
http://fahwiki.net/index.php/The_finstall_script
_ikki_
Posts: 27
Joined: Wed Dec 05, 2007 8:38 am

Post by _ikki_ »

Just a comment on the finstall backup utilly : regarding the source code, you make only 2 backup (every X minutes, 30 per proposed), ersasing the previous data (lesser than 1 hour)

For the stable client, there's no need to make a lot of backup, but for beta testing, it would be interesting to have a backup of a workunit at a given moment, even if the number of backup grow up (let the beta tester manually delete the archives periodically or make only 10 or 15 backups, according to the delay).

With this functionnality, il will be easy to report a bug in the V6 beta client for those who want to contribute in the project :)
Post Reply