ceph mon_clock_drift_allowed invalid

[huky]

I have set this in /etc/pve/ceph.conf
mon_clock_drift_allowed = 3 mon_clock_drift_warn_backoff = 30

and it is valid in config
ceph --admin-daemon /run/ceph/ceph-mon.node009.asok config show | grep clock "mon_clock_drift_allowed": "3.000000", "mon_clock_drift_warn_backoff": "30.000000", ......

but ceph still told me with "MAX 0.05s"
2021-04-28 09:02:39.131151 mon.node003 (mon.0) 2621051 : cluster [WRN] 1 clock skew 0.0871183s > max 0.05s 2021-04-28 09:06:39.131930 mon.node003 (mon.0) 2621134 : cluster [WRN] 1 clock skew 0.0906123s > max 0.05s 2021-04-28 09:11:09.132641 mon.node003 (mon.0) 2621227 : cluster [WRN] 1 clock skew 0.0631894s > max 0.05s 2021-04-28 09:11:35.047408 mon.node003 (mon.0) 2621241 : cluster [WRN] 1 clock skew 0.059099s > max 0.05s 2021-04-28 09:11:35.055041 mon.node003 (mon.0) 2621246 : cluster [WRN] overall HEALTH_WARN clock skew detected on mon.node009 2021-04-28 09:11:56.217482 mon.node003 (mon.0) 7 : cluster [INF] Health check cleared: MON_CLOCK_SKEW (was: clock skew detected on mon.node009)

How could I adjust the MAX to 3s?

 

[t.lamprecht]

What do you use for NTP?

In general it'd be better to fix the reasons for clock skew. If you use systemd-timesync it could be better to switch to chrony or openntpd. For example, apt install chrony.