>In CSA, is their any possibility of getting a report
>based on jobs run and classified according to users.
>Kindly specify me how to go about it.
I may be misunderstanding the format you want, but the
csacrep command can be used to create a report like the following
(different options add different types of resource usage information):
PROJECT USER LOGIN JOB
NAME ID NAME ID
======== ======== ======== ==================
Unknown 0 root 0x3de6
Unknown 0 root 0x3ef8
Unknown 0 root 0x3f14
Unknown 0 root 0x3f30
Unknown 0 root 0x3f49
Unknown 10320 kohnke 0x3de6
Unknown 10320 kohnke 0x3ef8
Unknown 10320 kohnke 0x3f14
Unknown 10320 kohnke 0x3f30
Unknown 10320 kohnke 0x3f49
See the csacrep man page for options to add resource usage info
to the above report. The csacrep command needs input data in the
form of cacct binary records. These come from use of the csacon
command which uses a sorted pacct file as input.
Normally, the sorted pacct file (spacct) is deleted after the
daily accounting run is done, i.e. the csarun command. There are
four user exits within that script where a site can take control
to do whatever the sites needs. So, here's what I did:
1) created the file /usr/local/sbin/csa.fef as root, permissions 755
with the following contents:
#! /bin/sh
cd ${WTIME_DIR}
cp ${SPACCT} /tmp/spacct
2) manually ran the daily accounting reports as root:
/usr/local/sbin/csarun 2> /var/csa/nite/fd2log
3) the csa.fef script was executed as part of csarun, which saved the spacct
file
in /tmp
4) used the csacon command to consolidate the information by job id and user
id:
/usr/local/sbin/csacon -ju -s /tmp/spacct > /tmp/cacct
5) used the csacrep command to create a report sorted by job id and user id:
/usr/local/sbin/csacrep -huJ < /tmp/cacct
---
Marlys Kohnke Silicon Graphics Inc.
kohnke@xxxxxxx 655F Lone Oak Drive
(651)683-5324 Eagan, MN 55121
|