[summary] die easy ultra 10

From: Giardina Marina <wavexx_at_ngweb.it>
Date: Fri Aug 23 2002 - 16:43:58 EDT
After some time, I can now summarize what happened with my ultra 10.

The problem was: an ultra 10 machine (built under year 2000, 2x8gb
ide disk, 256mb of ram, UltraSparc 2i 440mhz) started to crash quite
randomly with a variable system load.

I tried to change the ram modules, disable the PCI ethernet card and
change a couple of things but nothing worked.

After having some infos through sunmanagers, I run the iscda.sh script
and identified the famous ecache parity error. The cpu module should
be replaced.

The machine was two months out of warrany and I needed this machine
up&running in at maximum a couple of days.

I decided to try everything before sending it outside so I opened
the case and sequentially:

 - removed the cpu module
 - cleaned everything as best as I can with compressed air
 - reinserted the cpu module a couple of times to rougly test
   the connectors and displace a bit of dust.

After doing that, I've closed everything, fsck'ed the filesystem
again and leaved it working. By now, the machine has 75 days of
uptime with full load, with no errors.

Before cleaning, the machine was perfectly clean, so I'm thinking
that the cpu module was sightly displaced for some reasons
(maybe the high temperature during summer) and simply 
removing/reinserting it was enough.

Thanks to everyone that pointed me to iscda.sh and others
who replied. Sorry for the >long< time.
sunmanagers mailing list
Received on Fri Aug 23 16:46:07 2002

This archive was generated by hypermail 2.1.8 : Thu Mar 03 2016 - 06:42:52 EST