SUMMARY(2): Sun Fire x2270 AHCI watchdog timeout

[Bertold Kolics]

SUMMARY(2):
This is a follow-up to my message I sent 12/17/2009. I have been working
with Sun support every since. Sun suggested disabling CPU power management.
Unfortunately, this did not resolve the issue.

The only workaround I have found so far was to downgrade to Solaris 10 5/09.
I have been running several x2270 systems using Solaris 10 5/09 without any
hangs for several weeks.

Bertold


ORIGINAL ISSUE:
I have Sun Fire x2270 system running Solaris 10 10/09 and using 4 internal
SATA disks. The disks in this system are mirrored 4-way using ZFS. The
system locks up every 2-3 days. When this happens, I can't login from the
console (I can only enter the login name, but I never get to the password
prompt).

After power cycling the server,
- fmdump does not show any errors,
- /var/crash/<hostname> is empty,
- ZFS utilities don't indicate any disk errors,
- the service processor's event log has no relevant records,
- and the below lines can be seen in /var/adm/messages (i.e. these are the
last messages in the log before the reboot):

Dec  5 22:55:56 x2270-17 ahci: [ID 517647 kern.warning] WARNING: ahci0:
watchdog port 2 satapkt 0xfffffe9a07ea2540 timed out
Dec  5 22:56:11 x2270-17 ahci: [ID 517647 kern.warning] WARNING: ahci0:
watchdog port 0 satapkt 0xfffffe9a07ea21c0 timed out
Dec  5 22:56:11 x2270-17 ahci: [ID 517647 kern.warning] WARNING: ahci0:
watchdog port 0 satapkt 0xfffffe9a07ea0380 timed out
Dec  5 22:56:11 x2270-17 ahci: [ID 517647 kern.warning] WARNING: ahci0:
watchdog port 1 satapkt 0xfffffe9a07e71b60 timed out
Dec  5 22:56:11 x2270-17 ahci: [ID 517647 kern.warning] WARNING: ahci0:
watchdog port 1 satapkt 0xffffffffaec87b68 timed out
Dec  5 22:56:11 x2270-17 ahci: [ID 517647 kern.warning] WARNING: ahci0:
watchdog port 1 satapkt 0xfffffe996cdb4e08 timed out
Dec  5 22:56:11 x2270-17 ahci: [ID 517647 kern.warning] WARNING: ahci0:
watchdog port 1 satapkt 0xfffffe996cdbac48 timed out
Dec  5 22:56:11 x2270-17 ahci: [ID 517647 kern.warning] WARNING: ahci0:
watchdog port 3 satapkt 0xfffffe996ec23b60 timed out
Dec  5 22:56:11 x2270-17 ahci: [ID 517647 kern.warning] WARNING: ahci0:
watchdog port 3 satapkt 0xfffffe9a07e56d28 timed out

The system is on the latest firmware/BIOS/service processor release
available from Sun.

-- 
Bertold Kolics <bertold.kolics at unboundid.com>
Phone: +1.512.904.9130 x106, Fax: +1.512.519.4352