I am experiencing a funny situation on one of our machines.
Setup:
HP NetServer LX Pro (4 Cpus)
Kernel 2.2.1
DPT Raid 3334UW + 12x9Gig disks formed as 6 raid-1
512 Mb ram
128 Mb swap
Situation:
Each evening, at 23.00, 23.20 and 23.40, 3 cron jobs are started. They
recompiles the days work of Cobol-code (2100 objects) in 3 versions.
Every since 2.1.108 (maybe 2.1.107), the finish time of these 3 jobs has
increased - the number of objects are more or less the same. This is
the start/end time for one of the jobs:
Start End
Sun Feb 21 23:40:00 CET 1999 Mon Feb 22 02:54:53 CET 1999
Mon Feb 22 23:40:01 CET 1999 Tue Feb 23 06:17:14 CET 1999
Tue Feb 23 23:40:00 CET 1999 Wed Feb 24 07:04:27 CET 1999
Wed Feb 24 23:40:00 CET 1999 Thu Feb 25 07:23:21 CET 1999
Thu Feb 25 23:40:01 CET 1999 Fri Feb 26 08:59:14 CET 1999
As you can see, the end time spans from 02:54 to 08:59, without any
incease in objects. During 5 days, the system has become 6 hours
'slower'. The load seems to be rising faster and faster during this
period (This system is a lightly loaded system: 6-7 people is using it
at most times).
Going back to 2.1.108 gives us 'end'-times at 2:55 to 3:15 ....
What parameters can I play around with, to get back my normal
'end'-times?
Kind regards,
Brian
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/