W2100Z constantly shuts down
One of two W2100Z dual 2.4GHz Opteron workstations we have has started to shutdown almost immediately upon booting. Under Fedora Core 4 linux this machine reports that the Crtiical Temperature has been reached of 68C. This is despite the fact that when it does run normally I see from /proc/acpi that the temperature is measured at 44C. The Java Desktop support group told me today this is a known issue and that the earlier BIOS releases could actually damage the fans. However I doubt this in our case since the fans always seem to run and never surge. I also found that this shutdown occured when we booted off of the Supplemental 2.1 cd (making it impossible to flash the firmware). Has anyone else run into this problem? The fans do not sound abnormal when these shutdowns occur and the amber light remains on after the shutdown. We are having Sun service out to look at this but are baffled as to which components are likely at fault.
Jack
[979 byte] By [
] at [2007-11-25 22:38:31]
# 1
Looking in the DMI Event Log this morning, the problem seems clearer. I have a slew of 'system rear fan1 fault' errors. Unfortunately I can't stay booted up long enough to safely flash the bios so I'll have to wait for a replacement rear system fan. I am curious is this problem is due to a bad batch of system fans or if the tolerance on the BIOS fan control is just way too sensitive?
Jack
at 2007-7-5 14:09:03 >
# 2
After a few painful hours we managed to get the shutdown problem resolved. Both W2100Z workstations we have had Rev 1 fans and early BIOS firmware. We managed to get the machine which first exhibited the problem booted by changing our fan mode to four speed. That let the machine stay booted so that we could flash the new v2.2 firmware that came out yesterday. Afterwards, the machine became immune to the automatic shutdowns (which were logged in the DMI Event Log as 'system rear fan1 fault' previously. The odd part is that when we had this problem on that machine every 20th time we saw an unrecoverable memory error(...on different DIMMs). The second machine developed the problem immediately after we fixed the first (most likely due to the fact that it hadn't been power cycled in months). We were finally able to get that machine flashed by using the other machines system fan (which seemed to be a little better than the other fan). After flashing the firmware both machines seem immune now to the shutdons although we still will be replacing the Rev 1 fans with Rev 3 models.
Jack
at 2007-7-5 14:09:03 >
# 3
My w2100z is exhibiting this same behavior. The workstation was fine for over a year but then It was shipped and upon arrival it constantly reboots after being up for less than an hour. Nothing is being logged in the DMI event log. I flashed to the new bios on supp. CD 2.4 still having the issue.
Whats the next step in debugging this issue? Where do I order these new fans?
spqr at 2007-7-5 14:09:03 >
# 5
I have the ISO, but I can't get the system to stay up long enough to boot off the CD.I have the same errors (about the fan) in my log, but interestingly, the fan powers up fine.Is there a way to get the system to ignore the errors long enough to flash the BIOS?
