0

CentOS 5.x

My CentOS server appears to have jumped ahead 7 months. I'm not sure how or why. Are there any logs in CentOS that could provide more insight here?

Unfortunately, ntpd was not running. I've since configured/enabled it and the time/date has been corrected.

-M

Mike B
  • 11,871
  • 42
  • 107
  • 168

3 Answers3

1

You can use hwclock -r to read the time stored in the NVRAM. (I still like to call it CMOS, misnomer or not.) In all likelihood, you rebooted and the time is set incorrectly there.

If your mainboard battery is dead, you shutdown for 7 hours, and booted back up; you would lose that time.

Otherwise, it's potentially a time skew based compounded from a dying mainboard battery.

Warner
  • 23,756
  • 2
  • 59
  • 69
1

This is most defiantly a main board clock battery issue. If you are hosting it yourself then simply power down your machine, unplug it from its power source; remove the battery and check it with a battery checker. If its dead you can easily replace it; if not, Replace it making sure that it firmly connected and not loose, Then when booting reset the CMOS time to the current time/date.

However, If you are not hosting the server, Contact your hosts.

Patrick S
  • 11
  • 2
  • Actually, wouldn't it be possible if you're careful enough to remove it while running? The battery only powers the thing when there is no motherboard power afaik – Joris Sep 07 '10 at 17:59
1

On a completely different track, if this is a virtual rather than physical machine, there can be some problems introduced by the hosting machine. VMWare has a whitepaper out thereon this issue. I know I've personally come across it when I had an overloaded guest with nowhere near enough RAM. It wound up losing time and drifting horribly out of sync.

I'm well aware you didn't specifically mention virtualization, but this might be helpful info for anyone else who stumbles across this question.


--Christopher Karel

Christopher Karel
  • 6,582
  • 1
  • 28
  • 34