2

I am having problem with my dovecot on CentOS machine.

The error message is as below.

server dovecot: dovecot: Time just moved backwards by 1 seconds. I'll sleep now until we're back in present. http://wiki.dovecot.org/TimeMovedBackwards

I checked the link provided in the error message, I don't have ntpd running(did ps -ef | grep ntpd).

I found one pattern which would suggest that this happens at 4:36/37 (AM and PM).

maillog.2:Nov  1 16:36:59 server dovecot: dovecot: Time just moved backwards by 1 seconds. ...
maillog.2:Nov  3 04:37:00 server dovecot: dovecot: Time just moved backwards by 1 seconds. ...
maillog.2:Nov  4 04:37:00 server dovecot: dovecot: Time just moved backwards by 1 seconds. ...
maillog.2:Nov  5 04:36:59 server dovecot: dovecot: Time just moved backwards by 1 seconds. ...
maillog.2:Nov  6 16:36:59 server dovecot: dovecot: Time just moved backwards by 1 seconds. ...
maillog.3:Oct 26 16:36:59 server dovecot: dovecot: Time just moved backwards by 1 seconds. ...
maillog.3:Oct 29 04:36:57 server dovecot: dovecot: Time just moved backwards by 4 seconds. ...
maillog.3:Oct 29 16:37:00 server dovecot: dovecot: Time just moved backwards by 1 seconds. ...
maillog.3:Oct 30 16:37:00 server dovecot: dovecot: Time just moved backwards by 1 seconds. ...
maillog.3:Oct 31 16:37:00 server dovecot: dovecot: Time just moved backwards by 1 seconds. ...
maillog.4:Oct 20 04:36:59 server dovecot: dovecot: Time just moved backwards by 1 seconds. ...
maillog.4:Oct 21 16:37:00 server dovecot: dovecot: Time just moved backwards by 1 seconds. ...
maillog.4:Oct 22 16:36:59 server dovecot: dovecot: Time just moved backwards by 2 seconds. ...
maillog.4:Oct 24 16:36:57 server dovecot: dovecot: Time just moved backwards by 4 seconds. ...

I checked if any cron that does the time correction. I checked log at this time in /var/log/messsages but could not find anything specific.

Please suggest me a way to find out what causes this time shift. Also let me know if any other info required.

dfc
  • 1,341
  • 8
  • 16
bitkot
  • 131
  • 5
  • Is this really a server? Or just a VPS? If the latter, what technology (or worse yet, is it containerised)? Do you have anything in any of the (many) root cron files that might indicate what goes on at `37 4,16`? – MadHatter Nov 17 '15 at 09:11
  • @MadHatter Thanks, I currently don't have that information. I will edit the question once I get it. Currently I have added a cron entry which checks dovecote process every 5 minutes and start if not running as suggested by the wiki link with the error. – bitkot Nov 17 '15 at 09:22
  • I listed all the crons using `http://stackoverflow.com/a/134944/1629362` but I don't see any crons running at 4/16:36 – bitkot Nov 17 '15 at 09:24
  • You can find out what it is by running `virt-what` as root (you can `yum install virt-what` if it is not already installed). – Michael Hampton Nov 17 '15 at 09:30
  • I am not able to install `virt-what`, It showed `Error: Cannot find a valid baseurl for repo: update`. I followed [some steps](http://unix.stackexchange.com/questions/82566/yum-returns-errors-when-using) to correct that, it works now but still says `virt-what` not found when I do `yum install virt-what`. This is probably because the OS version is `CentOS 4.9` which is really old. – bitkot Nov 17 '15 at 12:16
  • 2
    That is too old; it shouldn't be running anymore! – Michael Hampton Nov 19 '15 at 18:05

1 Answers1

0

I'm going to lob an answer out there: It's virtualized, its clock is all over the place, and occasionally the clock is dragged back a whole second; probably after some load.

I think the error's benign, but I completely support @michael and his suggestion that this VM needs to be retired with great speed. It's really challenging to find the right words for polite conversation to express just how important it is that this host be retired tomorrow.

user2066657
  • 336
  • 2
  • 13