I have a quad AMD64 system (Tyan S4882UG2NR-D motherboard) with 16GB RAM
running RHEL4 update 1 (stock 2.6.9-11.ELsmp kernel).
I installed a new Adaptec 58300 SAS controller card connected to an Xtore
XJ-SA12-316R external chassis housing 12 Seagate ST375064 disks.
I downloaded and installed Adaptec adp94xx driver module (version
adp94xx-1.2.5918-rhel4, rpm package adp94xx.rhel4qu1.x86_64.rpm).
modprobe adp94xx caused all 12 drives to be attached and appear in
/dev/sdc,d,e etc. Each drive was partitioned using parted (gpt table) as
single partition and slaved together as 11 disk raid 5 array with 1 hot spare
using mdadm. Array re-sync was proceeding normally.
After approx 1 hour error messages concerning one of the disks appeared. 15
min later same error again, but this time more messages followed by HARD CRASH
of the system. Kernel panic type stuff on console.
Relevant part of /var/log/messages output attached below.
Adaptec suggested I upgrade card BIOS to latest which I did but did not help.
Since then I put the card in a dual AMD64 machine running the exact same OS
and I get the same "hr_ioreq_timedout" error.
I also managed to provoke the error by doing "cat /dev/sdc1 > /dev/null" on
all the disks in parallel so the problem is not specific to md.
Any help or info regarding similar experiences much appreciated.