Hi all,

I've got an issue where the time displayed by either timesync OR the xntpd logger screen is exactly one hour behind the time shown by "time" at the console prompt (which is correct).

Our NW6.5SP8 servers were changed from timesync to xntpd. No problems at the time - about a month ago.

Issue today with an NW6.5 SP8 box primarily used as a print server.

  • xntpd/timesync/the console all think they're in sync, no errors.
  • DS is happily synced.
  • It's using the same NTP sources (SLES10SP2) as other servers running a mix of timesync and xntpd.
  • App 1 (McAfee Netshield) has the same time as timesync/xntpd.
  • App2 (Moonwalk) has the same time as the console.
  • Prior to issue, server hex dumped, abend in apache.
  • Server came back up.

We had a similar issue on a different NW6.5SP8 box running GroupWise 7 where the following occurred:

  • Timesync was unloaded and xntpd was loaded.
  • timeserv.ncf was updated to load xntpd instead of timesync
  • Time all OK.
  • Later on (days) the server abended (backup s/w) and was booted.
  • Issue the same as above occurred - mail was being stamped 1hr behind - POA displayed wrong time, console was correct.
  • GW components stopped, config double-checked in GW, verified OK.
  • GW started up, same problem.
  • GW unloaded. Server reverted back to timesync from xntpd. Timesync one hour out. Console still OK.
  • GW load commands commented out in autoexec. Server rebooted gracefully.
  • Server comes up OK - time all OK
  • GW loaded - time OK.

The NTP sources (3 of them) are sound.

Does this mean that any time a server abends that it needs to be rebooted again after coming back up?!?
Am I missing something here?

Any suggestions appreciated!