since a recent update ( I think it's related to syslog-ng or systemd) my dmesg logs are filled with kernel timestamps even though I've had them disabled with print.time=N on my kernel line in grub for a very long time.
I've looked though syslog-ng.conf but I can't see anything related to time stamps.
I used a custom kernel and the kernel debug message PRINTK_TIME=n is set so I can't see what's turning on the timing messages. If a userspace tool is doing this, it shouldn't as I've disabled it at the kernel level.
What next to check?
Why wouldn't you suspect dmesg? use -t, --notime to get rid of the timestamps.
probably because I've disabled at kernel level the time stamps! why would a userspace tool default to turning them on? cat /var/log/boot shows them as having been logged.
cat /var/log/boot Fri Jul 20 19:21:28 2012: :: Adjusting system time and setting kernel timezone [BUSY] [DONE] Fri Jul 20 19:21:28 2012: :: Starting UDev Daemon [BUSY] [DONE] Fri Jul 20 19:21:28 2012: :: Triggering UDev uevents [BUSY] [DONE]
as do other logs.
I shouldn't have to use a switch that I've never had to use in the past. Time for a bug report.
Welcome to the ever changing world of open source. Linux 3.5 introduced structured logging, which dmesg now reads as of util-linux 2.22. By default, this includes timestamps which most people will find to be hugely beneficial in their logs for the purposes of troubleshooting.
I'm not using any switch with 'dmesg', but /var/log/boot indeed shows timestamps:
$ dmesg | tail [ 9.662139] EXT4-fs (sda3): re-mounted. Opts: (null) [ 9.776824] EXT4-fs (sda1): mounting ext2 file system using the ext4 subsystem [ 9.781108] EXT4-fs (sda1): mounted filesystem without journal. Opts: (null) [ 9.825871] EXT4-fs (sda4): mounting ext3 file system using the ext4 subsystem [ 9.951467] EXT4-fs (sda4): mounted filesystem with ordered data mode. Opts: (null) [ 9.987777] Adding 265068k swap on /dev/sda2. Priority:-1 extents:1 across:265068k [ 12.965433] Adding 255996k swap on /dev/zram0. Priority:60 extents:1 across:255996k SS [ 13.077880] IPv6: ADDRCONF(NETDEV_UP): eth0: link is not ready [ 13.080198] e100 0000:02:08.0: eth0: NIC Link is Up 100 Mbps Full Duplex [ 13.080751] IPv6: ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready $ tail /var/log/boot Sun Oct 14 12:56:26 2012: INIT: Entering runlevel: 3 Sun Oct 14 12:56:26 2012: :: Starting Syslog-NG [BUSY] [DONE] Sun Oct 14 12:56:27 2012: :: Starting D-BUS system messagebus [BUSY] [DONE] Sun Oct 14 12:56:28 2012: :: Starting acpid [BKGD] Sun Oct 14 12:56:28 2012: :: Starting network [BKGD] Sun Oct 14 12:56:28 2012: :: Starting crond [BKGD] Sun Oct 14 12:56:28 2012: :: Starting alsa [BKGD] Sun Oct 14 12:56:28 2012: :: Starting vnstat [BKGD] Sun Oct 14 12:56:28 2012: Setting up swapspace version 1, size = 255996 KiB Sun Oct 14 12:56:28 2012: no label, UUID=467ff282-5acf-4367-9e8e-76e4ed35361a
BTW, is there a way to localize the timestamps the way I can do with 'date':
$ LC_TIME=C date Sun Oct 14 15:50:34 CEST 2012 $ date nie, 14 paź 2012, 15:50:35 CEST
I'd love to find a way to display the timestamps the way that makes sense to me.