My Fit-PC2 can't keep the time
Posted: Wed Apr 21, 2010 7:50 pm
I'm having serious trouble with time keeping on my machine (using Debian testing). I'm not talking about a slow drift: it can easily lose more than 30min per hour. Here is an example session:
I do not see anything relevant in the syslog files. Dmesg sometimes gives me tons of "set_rtc_mmss: can't update from NN to MM", but not always.
The weird thing is that this exact same Debian installation was working perfectly fine at the very same location (tho on different hardware: I just took out the drive, replaced the big tower with the tiny Fit-PC2 and replugged the disk). So my suspicion is on the Fit-PC2.
Has someone seen something like that?
According to the "sleep 60", the loop should run about once per minute, but it sometimes ends up waiting five minutes or more. I'm running ntpd on this machine. Have also tried openntpd but it did not help.# (export TZ=UTC LANG=C; while sleep 60; do echo $(su - monnier -c "ssh fmsmemgm date"; hwclock; date); done)
Wed Apr 21 15:27:35 UTC 2010 Wed Apr 21 15:27:36 2010 -0.328802 seconds Wed Apr 21 15:27:35 UTC 2010
Wed Apr 21 15:30:35 UTC 2010 Wed Apr 21 15:30:36 2010 -0.172800 seconds Wed Apr 21 15:30:35 UTC 2010
Wed Apr 21 15:36:38 UTC 2010 Wed Apr 21 15:36:39 2010 -0.547803 seconds Wed Apr 21 15:31:38 UTC 2010
Wed Apr 21 15:52:41 UTC 2010 Wed Apr 21 15:52:42 2010 -0.703776 seconds Wed Apr 21 15:32:41 UTC 2010
Wed Apr 21 15:53:44 UTC 2010 Wed Apr 21 15:53:45 2010 -0.594683 seconds Wed Apr 21 15:33:44 UTC 2010
Wed Apr 21 15:59:47 UTC 2010 Wed Apr 21 15:59:48 2010 -0.641850 seconds Wed Apr 21 15:34:47 UTC 2010
Wed Apr 21 16:00:49 UTC 2010 Wed Apr 21 16:00:51 2010 -0.813433 seconds Wed Apr 21 15:35:50 UTC 2010
Wed Apr 21 16:11:52 UTC 2010 Wed Apr 21 16:11:54 2010 -0.985030 seconds Wed Apr 21 15:36:53 UTC 2010
#
I do not see anything relevant in the syslog files. Dmesg sometimes gives me tons of "set_rtc_mmss: can't update from NN to MM", but not always.
The weird thing is that this exact same Debian installation was working perfectly fine at the very same location (tho on different hardware: I just took out the drive, replaced the big tower with the tiny Fit-PC2 and replugged the disk). So my suspicion is on the Fit-PC2.
Has someone seen something like that?