I just installed a new system containing an Intel Endeavour with a P133,
32 MB RAM (60ns) and a Quantum EIDE drive (2.5 GB). I use Linux 1.2.13.
The IDE stuff just doesn't seem to work properly:
After some disk I/O (some newsgroups are received, but other I/O, like many
tar/untar processes, causes the same problems), the filesystems are
corrupted. Always. The system then breaks down between 30 minutes
and 6 hours or so.
When I reboot the system in the normal way, after about 10 minutes of
heavy disk I/O, and force a filesystem check in single user mode,
the filesystem is _always_ corrupted (and this is after a clean unmount!).
In a lot of cases, the kernel even panics or hangs when doing e2fsck then!
Also, during e2fsck, I often get the message "VFS: Wrong blocksize
on device 3/2".
And now the interesting point: I also have an ASUS P55T2P4 (the new one),
a P166, and the same RAM and disk (that is, same brand, specs, speed, etc),
and that works flawlessly (i'm doing the same tests, of course).
So, my questions:
- Is this maybe a known bug in the Intel Endeavour?
- Or is the Intel board triggering a bug in the 1.2.13 IDE driver?
- Is there a patch for the 1.2.13 IDE driver (for example) to solve
the problem or to act as a workaround?
- Will Linux 2.0 solve the problem (I'll probably try this myself,
but I just don't want to run 2.0 now on that particular system
at this moment)?
Thanks in advance for any help.
-- -- Jos Vos <jos@xos.nl> -- X/OS Experts in Open Systems BV | Phone: +31 20 6938364 -- Amsterdam, The Netherlands | Fax: +31 20 6948204