Interesting. I have a system that used to run SCO. Its a 90MHz Pentium
with PCI and an Adaptec 2940 controller. It worked with SCO but with
Linux is was really flakey (data wasn't getting written to the SCSI
HD so lots of corrupt files). I tried a 1542 in place of the 2940 and I
thought at first all was well. But I've noticed a block of NULLs in the
/var/log/messages file which makes me think it is still sometime failing
on writes (write is done to /var/log/messages, data doesn't actually
get written but file position is updated, more data written and it does
get to the disk, now you have a hole in the file that shows up as NULLs).
Anyone have an idea on how to fix this? Are there some known chipsets
that don't agree with Linux? Worked fine with SCO before this...
I think its a Micronics motherboard and I'll be able to check the
motherboard on Tuesday to see just what chips it has. I could try
turning off the cache (thats what I often hear people having trouble
with) but that will really slow it down and is unacceptable. Also I
think it sounds more like a DMA problem than cache.
I would **LOVE** if someone had any clue on this!!! I stuck my neck out
to replace SCO with Linux and its been kinda chopped off as the Linux
ended up much faster but flakey while the SCO was slow but sure... If
I can't get this solved its either back to SCO (ick!) or I have to
convince everyone that we have to fork out cash for a new motherboard.
> Tim Hedges
>
Hope someone can help me!
======================================================================
Brad Pepers Proud supporter of Linux and
Ramparts Management Group Ltd. Caldera in Canada!
ramparts@agt.net
http://www.agt.net/public/ramparts Linux rules!