Debian 3.0r1 testing with an occasional unstable (new libc) libc6 2.3.1-14 ntp 4.1.0-8 sysklogd 1.4.1-10
System runs ntpd and is locked on hard to correct time. Ntp is not reporting any time jumps, loss of lock, etc. Hwclock is UTC and correctly synced with the system time. Every once in a while, I'd see an item in the mailq (exim) with a queue time of -64 minutes (or so). So I started doing some snooping in the log files. There are entries in the log that are incorrect as well. (Example from syslog below). This is not a case of things being written from a cache out of order (but the data actually being correct). Mar 19 02:09:12 bambusa identd[20722]: started Mar 19 02:14:50 bambusa identd[20746]: started Mar 19 02:19:20 bambusa identd[20787]: started Mar 19 03:35:59 bambusa identd[20821]: started Mar 19 02:32:55 bambusa identd[20858]: started Mar 19 02:42:44 bambusa identd[20930]: started Mar 19 02:45:25 bambusa identd[20949]: started $TZ = "" bambusa:~/logcheck# date Thu Mar 20 17:21:35 CST 2003 bambusa:~/logcheck# date -u Thu Mar 20 23:21:39 UTC 2003 bambusa:~/logcheck# date -R Thu, 20 Mar 2003 17:21:42 -0600 These appear to be correct. The error is always approximately 70 +/- 10 minutes. The error shows up in exim, sysklog, and a couple of other places, so I assume libc. Any idea where to start looking???? I checked bugs.debian.org under libc6, but I didn't see anything that appeared appropriate. -Jeff -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
