Messages in this thread | | | Date | Mon, 15 Mar 1999 09:42:25 -0600 | From | Jon Mitchell <> | Subject | system SLOOOWWW down on 2.0.37-ac5 SMP |
| |
We are running a thirty user terminal (ncurses) based application that uses fairly large ISAM files (300-500 MB) and a fair amount of memory, files open, locks open, etc per user. We are running this on a Compaq Proliant 1600 (tlan net drivers, two Pent II 400's, ncr83c5X9 driver for our external tape drive) with a DPT 3334UW controller with a RAID 5 array on three disks. The box is a Redhat-5.2 box with the latest patches running kernel 2.0.37-ac5 with SMP = 1. When it boots up both CPU's report a normal bogomips reading and appear to be functioning properly.
Problem: Between 2-4 days after rebooting when the employees come in the morning the system is incredibly slow. Using top it shows almost all the load is system load (about 95%). We are forced to get everyone to log off and reboot. Even with all users logged off the system load is incredibly high. Then we kill our job scheduler (something like cron), still incredibly slow. We've tried killing a number of the normal tcp and other system services, etc... but it never recovers. Running shutdown takes about half an hour because it is so slow, and it never speeds up after any particular process is stopped in an rc script. We recompiled with 2.0.37-ac8 and turned off SMP on Friday and so far things are good (crossing fingers).
Is this a known problem? What other things can I look at? There are no relevant errors in any of the syslog files, no processes that seem to be out of control, etc, etc...
Thanks for any assistance,
Jon Mitchell Systems Engineer, Subject Wills and Company jrm@swc.com
- 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/
| |