Hi,
Am 11.03.2013 um 15:52 schrieb Txema Heredia Genestar:
> This morning I have realized that I do not always receive mails sent from my
> jobs. I have been toying around with this script:
>
> #!/bin/bash
> echo "sleep 1" | /bin/mail -s "sleep ${1}" ${LOGNAME} ;
> sleep 60;
> echo "awake 2" | /bin/mail -s "awake ${1}" ${LOGNAME} ;
> sleep 10;
> echo "test 3" | /bin/mail -s "test ${1}" ${LOGNAME} ;
> ###sleep 1;
The mail interface will fork at some point the maildrop process. When the job
finishes the remaining processes will be killed by intention by SGE - if it
happens too fast, the mail ist never send as you observed.
What you can try is something like:
echo "test 3" | /bin/mail -S sendwait -s "test ${1}" ${LOGNAME} ;
-- Reuti
> If I submit this script, I only receive the 2 first mails, but never the
> third one. But if I uncomment the last sleep line, I do receive the third
> email.
>
> I have been playing further around this and it seems that, If a mail is the
> last command issued by a job, or if it is followed by some "fast" command I
> still don't receive the 3rd mail. But if the last mail command is followed by
> some "longer" command , I do receive them. A few examples:
>
> No mail:
> - ls
> - ls > /dev/null
> - exit 0
> - sleep 0
> - hostname
> - hostname (2x) & (3x)
>
> Receive mail:
> - sleep 1
> - top -b -n 1
> - hostname (4+x)
>
> A few additional twists:
> - If I log into the node and run the script, I receive everything.
> - "ssh node script" I also receive all mails
> - If there is a epilog script, the epilog is able to send mails even if the
> job failed to send the 3rd one.
>
> Any ideas of what is causing this? Is this a bug of SGE or a
> missconfiguration of my system?
>
> Thanks,
>
> Txema
> _______________________________________________
> users mailing list
> [email protected]
> https://gridengine.org/mailman/listinfo/users
_______________________________________________
users mailing list
[email protected]
https://gridengine.org/mailman/listinfo/users
