I had a machine with RH 6.0+ (this machine works for 2 years on the same HW without any trouble) and I upgraded it to the 7.0 (kernel-2.2.16-22 and kernel-2.2.17-14 from updates). The installer decided to use i686 kernel and Glibc (as uname -m display). The system boot normally but after a while (say 15-30 minutes) when the system was runnig and there was some disk operations the partitions with ext2 filesystem (I have only ext2 FS) went to hell (was almost unusable, e2fsck did not help). I decided to use i386 libs, kernel (fresh installation). The system run for 12 or 16 hours and then the same as above - there was a problems with ext2 filesystem. I decided to use another HW (486 procesor on a different motherboard) and everything is ok again. This seems that kgcc produce bad code for Cyrix M II. I did not tested any of 2.4.x kernels as this HW is in another firm. Reproducible: Always, update or fresh install doesn't matter This is high severity bug but 2.2.x is obsolete and there could be some comment about Cyrix processors when 2.2.19 will go out.
This looks more consistent with hardware than software problems, especially as its one of the notorious Cyrix heating elements as a processor. Does the machine pass stress tests like memtest86 ?
It could be possible. I have no longer access to this machine so I can't test it. The server's HW has been replaced by another.
Just spotted this bug still open. Without reproduceability, there isn't much can be done, especially since it looks like a hardware issue likely. Closing bug as NOTABUG for now. If you reproduce again, and the machine passes a memtest86 test, feel free to reopen for further investigation.