[squid-users] sda + sdb + hda = trouble?

Hi

I have had a perfectly good squid box running for about 4 years I
started off with 1 X 4 gig seagate scsi (1999)

I added 1 X 9 gig seagate scsi (2000) by adding another cache_dir entry,
squid -z etc..etc.

This went faultlessly

Now I add in addition 1 X 80 gig Seagate ide and am getting strange
lockup problems after 25% full.

Can you please help me solve this??

Clues:
When I add the cache_dir line for the ide, restart, then do a top, the
mem(ram) usage goes ballistic to 97% and stays at 97%, also I can hear
one of the scsi drives working like crazy.

Then I comment the cache_dir line for the ide out, restart, and it loads
up fine, like normal ie: cpu and mem usage up for about a minute then
drops off to wait for a request

I did find a few bad journal entries (sdb1 and hda1) when I did an fsck,
which stopped the crashing for a while but this didn't fix the ide/mem
usage problem. I don't know whether this bit has to do with the cause or
is an outcome of the ide/mem problem.

SYSTEM = Aopen mx5s, 900 celeron, 256 ram, 256 swap, Redhat 7.3,
squid-2.4.STABLE6-6.7.3.i386.rpm, 4gig scsi, 9 gig scsi, 80gig ATA100
ide, all ext3.

Kind regards

Stuart Clark
Received on Mon Jan 06 2003 - 05:20:55 MST