Server Time slows down

[karlos]

Running VE 1.5 on an Intel Quad core, After a reboot the server time is correct, then it slowly slows down, something like 45 minutes slower over a week. A reboot fix's it.

I noticed these kernal errors also(maybe to do with vzdump),

kjournald starting. Commit interval 5 seconds
EXT3 FS on dm-3: internal journal
EXT3-fs: dm-3 16 orphan inodes deleted
EXT3-fs: recovery complete.
EXT3-fs: mounted filesystem with ordered data mode

Can anyone point me in the right direction on the potential cause?

running kernel: 2.6.18-2-pve
proxmox-ve-2.6.18: 1.5-5
pve-kernel-2.6.18-2-pve: 2.6.18-5
qemu-server: 1.1-11
pve-firmware: 1.0-3
libpve-storage-perl: 1.0-8
vncterm: 0.9-2
vzctl: 3.0.23-1pve8
vzdump: 1.2-5
vzprocps: 2.0.11-1dso2
vzquota: 3.0.11-1
pve-qemu-kvm-2.6.18: 0.9.1-5

 

[tom]

you talk about the host? ntp working?

or a guest system?

 

[karlos]

Yes, the proxmox time is wrong, upon reboot it is fine but slows down again. I'm not familiar with ntp...

 

[karlos]

Has anyone any sugestions why this may be happening? Is there a way to re adjust the time perhaps?

 

[tom]

a good start of learning ntp is reading the manual pages:

man ntpd

 

[karlos]

ok will do, do you think the' orphan inode deleted' has any bearing on the problem?

edit: I restarted ntp and it reset to the correct time but its not a permanent fix i imagine.

 

[alain]

Hi karlos,

It is not sufficient to install the ntp service. You have to configure it in /etc/ntp.conf. In default configuration, you have :
.....
# You do need to talk to an NTP server or two (or three).
#server ntp.your-provider.example
....

Provide here the adddress of a valid NTP server that is available on your network (there should be one in any network, as synchronize all machines in a network is very imortant, for example for authentication, but also for Proxmox cluster synchronisation).
There are NTP servers on the Net, but often they are not reachable, due to firewall policies.

You shoul also install the ntpdate utility :
# apt-get install ntpdate

It permits to synchronize time by issuing a command :
# /etc/init.d/ntp stop
# ntpdate my_ntpserver.domain.com
# /etc/init.d/ntp start

Alain

 

[karlos]

cool, thanks Alain, but what is the cause of loosing time this quickly? I can't see why it would loose this much time...just because of potentially not finding an ntp server...everything else on the network has no problem with ntp...

this appears to be the default in ntp.conf so can i just change this but keep the same format?:
server 0.debian.pool.ntp.org iburst dynamic

eg: server 0.time.windows.com iburst dynamic

 

[alain]

I don't know why your server is loosing time so quickly, perhaps a battery or frequency problem of your clock on the motherboard ?

For the NTP server, test with npdate if it is indeed reachable...

Alain

 

[karlos]

I don't know how to test that...I just had another error type, perhaps after the ntp restart -

time.c can't update cmos clock from 0-59