| Author |
 Topic  |
|
|
Adria
New Member
19 Posts |
 Posted - Nov 10 2011 : 4:03:19 PM
|
| I am running a job that is taking a while to finish. Is there a way I can have the cluster notify me via email when the job is complete? |
|
|
petty
BIAC Staff
USA
453 Posts |
Posted - Nov 10 2011 : 4:08:27 PM
|
you can add a notification if you didn't do it at submission time:
qalter -m e -M email@address.com JOBID
-m is the email notication flag 'e' means ending |
 |
|
|
Adria
New Member
19 Posts |
Posted - Nov 10 2011 : 5:09:46 PM
|
| Thank you! Now what if I have a job running, but it doesn't finish before I leave work. If I log out of my computer, and return tomorrow, will it start off where it ended the day before, or will it rerun the whole job? |
|
|
|
petty
BIAC Staff
USA
453 Posts |
Posted - Nov 10 2011 : 8:14:36 PM
|
If its a batch job it will keep running, those are not tied to your computer at all.
It it's interactive it will die when your shell closes. |
|
|
|
Adria
New Member
19 Posts |
Posted - Nov 17 2011 : 3:07:52 PM
|
| Hi, thanks for the help. I have another question. I have set up a qsub script for FSLVBM, but I don't know when and where to enter the command to send a notification to my email address. I have been referencing the conversation between you and user dit11: http://www.biac.duke.edu/forums/topic.asp?TOPIC_ID=1491&SearchTerms=qsub |
|
|
|
syam.gadde
BIAC Staff
USA
421 Posts |
|
|
Adria
New Member
19 Posts |
Posted - Nov 17 2011 : 3:35:49 PM
|
| So does that mean that qsub will automatically send me an email without me having to send out an option, or are there command line options I have to use? |
|
|
|
petty
BIAC Staff
USA
453 Posts |
Posted - Nov 17 2011 : 4:00:25 PM
|
Put these two things in your qsub script
#$ -m ea
#$ -M user@school.edu
"-m ea" means to email you if the job ends or is altered
"-M " is the email address to send the notification too
you'll need both arguments |
|
|
|
Adria
New Member
19 Posts |
Posted - Nov 17 2011 : 4:14:52 PM
|
I typed in the -M command and the job went through, but I got an error saying "[akv7@node51 data]$ qsub -v EXPERIMENTS=fse.01 fslvbm.sh
Invalid EMAIL detected!
searching users for : akv7
changing email to adria.vidovic@duke.edu
Your job 2834398 ("fslvbm.sh") has been submitted
[akv7@node51 data]$ more fslvbm.sh
#!/bin/sh"
When I went back to check to see if I put in the right email address, it turns out that I typed it correctly. What do you think is happening?
|
|
|
|
petty
BIAC Staff
USA
453 Posts |
Posted - Nov 17 2011 : 4:35:36 PM
|
2 things: You need the "#" symbol infront of all of your SGE flags inside the script:
# -- END USER DIRECTIVE --
$ -M adria.vidovic@duke.edu
# -- BEGIN USER SCRIPT --
needs to be
# -- END USER DIRECTIVE --
#$ -M adria.vidovic@duke.edu
# -- BEGIN USER SCRIPT --
and you need to delete:
#$ -M someone@somewhere.edu
|
|
|
|
Adria
New Member
19 Posts |
Posted - Nov 17 2011 : 4:46:20 PM
|
| Ok great. Thank you! |
|
|
|
Adria
New Member
19 Posts |
Posted - Nov 17 2011 : 5:10:24 PM
|
| Hi, Just for future reference, what does SGE stand for? |
|
|
|
petty
BIAC Staff
USA
453 Posts |
Posted - Nov 17 2011 : 6:21:56 PM
|
| Sun grid engine |
|
|
|
Adria
New Member
19 Posts |
Posted - Dec 05 2011 : 09:24:36 AM
|
I received this error in the email I had BIAC send me.
Job 2834410 (fslvbm.sh) Aborted
Exit Status = -1
Signal = unknown signal
User = akv7
Queue = users.q@node56.local
Host = node56.local
Start Time = <unknown>
End Time = <unknown>
CPU = NA
Max vmem = NA
failed before writing exit_status because:
shepherd exited with exit status 19: before writing exit_status
What does this mean and is there anything I need to change in my script to make it work? Thank you!
|
|
|
|
petty
BIAC Staff
USA
453 Posts |
Posted - Dec 05 2011 : 09:34:39 AM
|
looks like if failed before it could even make it to a node, which is why all those values are blank.
you can try "dos2unix" on the script to remove any invalid characters, otherwise its likely a scripting error somewhere early in the script
|
|
|
| |
Topic |
|
Job status email notification
