[PVE-User] HEALTH_WARN clock skew detected on mon.1

[Mark Schouten]

On Mon, 2017-03-13 at 16:13 +0100, lists wrote:
> Occasionally (three times a month or so) we are getting "clock skew 
> detected" errors on our 3-host proxmox 4.3-1/e7cdc165 cluster.

Ceph is pretty critical about timesync.

Probably one of your hardwareclocks is a bit flaky, which causes things
to skew more than 0.5 seconds. I've 'fixed' this by doing 
minpoll 4 maxpoll 6

Which causes ntp to check more often (instead of every 1024 seconds).

-- 
Kerio Operator in de Cloud? https://www.kerioindecloud.nl/
Mark Schouten  | Tuxis Internet Engineering
KvK: 61527076  | http://www.tuxis.nl/
T: 0318 200208 | info at tuxis.nl