Red Hat Bugzilla – Bug 220984
Kernel hang on Supermicro P4SCT with sata_mv
Last modified: 2007-11-30 17:07:39 EST
This is Supermicro P4SCT+ rev. 1.11 with latest BIOS update (1.2b, 06/12/2006).
The system has been reported to me as stable with FC3 in the past. Now with
RHEL4 or RHEL5b2 I see random hard hangs (after minutes to several hours of
uptime). There are 7 discs (SATA WD4000KD-00N) connected to the MV88SX5081
8-port SATA I PCI-X Controller. I tryed binary kernel module instead of sata_mv
too with no luck (there is RAID5 array but with no dmraid now).
I tryed to disable hyperthreading, update BIOS, boot with noapic kernel
parameter with no luck. The system hanged during startup sequence on hotplud and
smartd daemon start so I disabled these. Other hangs was after some time with no
traces or OOPSes in the logs or on the screen.
I still have no clue about what going on (maybe SATA driver/card?). I moved the
system to my home to be able to do more testing so please ask me for what I have
The system is now running the latest test kernel (2.6.18-1.2913.2.1.el5.gtest.3)
for RHEL5 from http://people.redhat.com/agospoda/rhel5/gtest/.
Created attachment 144565 [details]
dmesg output (kernel-2.6.18-1.2913.2.1.el5.gtest.3)
Created attachment 144566 [details]
Created attachment 144567 [details]
I'm using kernel 2.6.18-1.2962.2.1.el5.gtest.4.i686 (HT off) from
http://people.redhat.com/agospoda/#rhel5 and the system seems to be stable last
few days (it survives rebuilding RAID5 arrays). Also I put off console blanking
by "setterm -blank 0 > /dev/tty[1-8]".
I'm going to do some stress tests to make sure it will not hang again.
It seems like HW bug on motherboard (the last piece of HW that has been claimed).
I'm not able to close bug as NOTABUG even I'm a submiter. Please close. Thank you.
Closing at the request of reporter.