Home/ Questions/Q 8688319
In Process

The Archive Base Latest Questions

Editorial Team
  • 0
Asked: 2026-06-12T23:20:00+00:00

I’m having issues with a bash script. Basically, in my script I run the

  • 0

I’m having issues with a bash script. Basically, in my script I run the command

bjobs -u $1

$1 is the script argument, and it’s the id of a user on my cluster. If user $1 has no running/pending jobs on the cluster, bjobs prints

 No unfinished jobs found

to stdout. Also, exit status is 255. If user $1 does have jobs, bjobs print:

JOBID   USER    STAT  QUEUE      FROM_HOST   EXEC_HOST   JOB_NAME   SUBMIT_TIME
45823   .......
45824   .......
.
.

and exit status is 0. I would like awk to parse the output of bjobs. I tried 

bjobs -u $1 | awk '...'

where ‘…’ is my awk parsing program. This doesn’t work so well, because apparently when user $1 doesn’t have running/pending jobs, the string “No unfinished jobs found” is not passed to awk. The same happens with

bjobs -u $1 > foo
awk '...' foo

Any suggestions?

Share

Leave an answer

You must login to add an answer.

Need An Account,

1 Answer

  1. Editorial Team

    Maybe that message (being an error of sorts, I guess) is coming to STDERR not STDOUT.

    You can redirect IO streams in bash with > to file descriptors, and 2 is STDERR.

    Try:

    bjobs -u $1 2>&1 | awk '...'

    And see if that solves it for you. The 2>&1 part means redirect STDERR (2) to STDOUT (1), and STDOUT is what is piped into awk.

    More info: http://www.tldp.org/LDP/abs/html/io-redirection.html

    • 0