ceph health clock skew detected

admin

MON_CLOCK_SKEW: clock skew detected on mon.ch-mon-1, mon.ch-mon-2
- l2 \5 X4 a' |5 Uceph -s                                                                       Fri Sep 28 13:57:25 2018
  cluster:
    id:     e6d7621d-4b97-4791-9d9b-92829a6f3044
    health: HEALTH_WARN
$ K5 G" ]+ d) ?9 E) a            clock skew detected on mon.ch-mon-1, mon.ch-mon-2
5 G! q7 k: G( m  x; g6 G, |) s  services:
    mon: 7 daemons, quorum ch-osd-1,ch-osd-2,ch-osd-3,ch-osd-4,ch-mon-1,ch-mon-2,ch-mon-3
$ r' D$ @% c2 g: V    mgr: ch-mon-1(active)
    osd: 12 osds: 12 up, 12 in
  data:
% _2 d# L$ J+ b( O% l    pools:   0 pools, 0 pgs
    objects: 0 objects, 0B
& h7 W" \7 C0 Q. x    usage:   12.1GiB used, 1.16TiB / 1.17TiB avail
    pgs:
8 k! P& z$ [- M' ^: T; F1 c7 n2 Y8 b

admin

ceph health detail
0 W. s9 c2 f/ l; o  gHEALTH_WARN clock skew detected on mon.ceph2, mon.ceph3, mon.ceph4
MON_CLOCK_SKEW clock skew detected on mon.ceph2, mon.ceph3, mon.ceph4
    mon.ceph2 addr 192.168.90.86:6789/0 clock skew 0.0630122s > max 0.05s (latency 0.00304483s)
    mon.ceph3 addr 192.168.90.87:6789/0 clock skew 0.076804s > max 0.05s (latency 0.00364158s)
! o  [% o: L* ]9 x% e) w/ M1 z- p6 H) l    mon.ceph4 addr 192.168.90.88:6789/0 clock skew 0.115598s > max 0.05s (latency 0.00468623s)
. S! N) n2 H2 u

解决办法是：
停掉服务ntp
1 s) a5 i4 ?6 Q5 W4 S! t2 H# systemctl stop ntpd.service
同步时间：
ntpdate -u 192.168.90.85
, s7 n8 {1 i1 C/ b+ F' k. P29 Sep 15:32:31 ntpdate[4396]: adjust time server 192.168.90.85 offset 0.034117 sec
, p+ {  D+ w' _- L# v  J9 r启动ntpd服务
systemctl start ntpd.service
' T6 C  j, Y1 B0 l
ceph health detail
  @3 A  H6 Z9 a; C0 Y' @2 l$ SHEALTH_WARN clock skew detected on mon.ceph3, mon.ceph4
MON_CLOCK_SKEW clock skew detected on mon.ceph3, mon.ceph4
    mon.ceph3 addr 192.168.90.87:6789/0 clock skew 0.0640733s > max 0.05s (latency 0.00336378s)
    mon.ceph4 addr 192.168.90.88:6789/0 clock skew 0.115242s > max 0.05s (latency 0.00423552s)

+ h9 I2 v7 z4 G6 ]) W

admin

systemctl restart ceph-mon.target
如果还是出现这样的问题，重启下ceph的mon服务
4 `3 [5 b: K- @  J$ Y$ ~ ceph health detail
0 U7 h+ L8 t) K- zHEALTH_OK
[root@ceph1 ~]# ceph -s
( |. X0 D0 \% G( Z3 M( O! p: ]% B  cluster:
    id:     9b8bcc5e-a0cf-4617-9e14-78acaf79e690
8 Y8 j6 F6 q2 A4 M, j1 m    health: HEALTH_OK

  services:
& ~  C) ?& [% L7 i    mon: 4 daemons, quorum ceph1,ceph2,ceph3,ceph4
    mgr: ceph1(active), standbys: ceph4, ceph2, ceph3
! E9 r% }9 H( j    osd: 8 osds: 8 up, 8 in
! H! T- b+ W2 j! E
; Y" w/ ?: X: L& X& W# z  data:
/ U& l0 A- R2 U; @1 [: m+ Z" H! [    pools:   5 pools, 448 pgs
    objects: 0 objects, 0B
  F1 }9 d0 S, m0 E% u    usage:   876MiB used, 799GiB / 800GiB avail
    pgs:     448 active+clean
  b  D, ?' X* E" G5 z7 o
5 O9 Z( x! r$ W3 R7 a
然后再启动ntpd
  @- m: E; ]4 k4 K) p% nsystemctl start ntpd.service