ntp (in a netvm) sudenly fails... (happend more than once)

[marmarek]

Reported by joanna on 29 Jan 2012 11:53 UTC

[user](user@netvm)$ sudo cat /var/log/messages | grep ntp | grep systemd
Jan 29 00:53:05 netvm systemd[ntpd.service: main process exited, code=exited, status=255
Jan 29 00:53:05 netvm systemd[1](1]:): Unit ntpd.service entered failed state.
[user](user@netvm)$ 

This results in all clocks in my system getting hugely desynced, by an offset of several hours (which is strange, BTW, because ntp crashed only last night, and this morning my clocks went crazy...)

Migrated-From: https://wiki.qubes-os.org/ticket/432

[marmarek]

Comment by joanna on 30 Jan 2012 11:16 UTC
Maybe we should exclude the "clock VM" from being updated by Dom0 upon resume, as it might be causing the ntpd failure?

[marmarek]

Comment by joanna on 30 Jan 2012 11:16 UTC
Maybe we should exclude the "clock VM" from being updated by Dom0 upon resume, as it might be causing the ntpd failure?

[marmarek]

Comment by marmarek on 30 Jan 2012 11:31 UTC
I think yes - clock VM is the source of time for dom0 (and in result other VMs).

[marmarek]

Comment by marmarek on 30 Jan 2012 11:31 UTC
I think yes - clock VM is the source of time for dom0 (and in result other VMs).

[marmarek]

Comment by joanna on 30 Jan 2012 12:10 UTC
Most likely fixed by this:

http://git.qubes-os.org/?p=joanna/core.git;a=commitdiff;h=ddecfab7e81df49d0ff1facf2eedd26713da522chttp://git.qubes-os.org/?p=joanna/core.git;a=commitdiff;h=ddecfab7e81df49d0ff1facf2eedd26713da522c

[marmarek]

Comment by joanna on 30 Jan 2012 12:10 UTC
Most likely fixed by this:

http://git.qubes-os.org/?p=joanna/core.git;a=commitdiff;h=ddecfab7e81df49d0ff1facf2eedd26713da522chttp://git.qubes-os.org/?p=joanna/core.git;a=commitdiff;h=ddecfab7e81df49d0ff1facf2eedd26713da522c