NTP issue in Virtual server

[dhaniklal]

Hi,

I am facing issue in ntp, it is getting killed frequently. see below error;
# /etc/init.d/ntpd status
ntpd dead but pid file exists

After starting service it is works fine but again after sometimes it get in dead postition.
Please help me or give me suggestion ?

Thanks,
Dhanik

[Roxoff]

Is there anything in the logs that suggests why the service terminated? Take a look in /var/logs/messages in particular, but the error could be in other files, such as /var/log/secure, so read around.

Are you running VMWare and using VMWare Tools to keep the clock in sync with the host machine? Could this be interfering and causing ntpd to stall?

[Irithori]

It is a long shot, as you did not provide too much detail.

But I know, that timekeeping in virtual machines is not as easy.
Roxoff: installation of the vmware tools helps, but is not enough from my experience.
The time inside the VMs will still fluctuate (is there such a word? ), due to for example CPU throttling on the host server.
And I think, ntp doesnt like it very much if the time slows down/speeds up.

One can set these configs to minimze the effect
- on the host: /etc/vmware/config add
host.cpukHz = 2330000 # 2.33Ghz = max Speed of one CPU
host.noTSC = TRUE
ptsc.noTSC = TRUE
tools.syncTime = "TRUE"
- in each VM:
* /etc/grub.conf: add clocksource=pit as a kernel option (on older kernels this options is called clock=pit)

Also read here:
http://www.vmware.com/pdf/vmware_timekeeping.pdf


We ultimately abandoned vmware server and went for ESXi, which is more robust in that context.

[coopstah13]

i don't have any problems using vmware tools to keep it synced to the host, which is in turn synced to the actual NTP server on our network

[Irithori]

coopstah13: Do your VMs run 24/7/365?

But on the other hand: my info is about 1,5y old.
It might just work recently

[coopstah13]

my vm's are up 24/7, they haven't reached 365 yet though, certainly up for months