Les Mikesell wrote:
Random errors would make me think it is a problem with interrupt
handling or DMA that you probably won't duplicate in a sequential
test. Have you checked for firmware updates?
I managed to get an install to fail without locking up last night,
(still using the controller, but with seperate disks, no raid) and that
gave me the opportunity to get an installer dump out of anaconda.
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=189570
The raid controller has the latest firmware. Motherboard has recent
firmware, not sure about the very latest, will check.
If you think that it may be something only bought on by random
read/write then I can run bonnie to simulate that, howver where/how
would I run it.
I could use a "livecd" type distro, but I've already proved that other
distros do not suffer from this problem, it does seem to be something
unique (at the moment) to FC5.
--
Cheers,
Dave Russell
