octopus版本ceph 分布式问题 ceph status HEALTH_WARN

admin

ceph status
1 h4 r* p$ X4 i1 D8 u3 b% J5 [  cluster:
. ?) l/ ?1 `; B    id:     ed60a380-c9bd-11eb-ad78-000c29ad02de
    health: HEALTH_WARN
) n  }* y1 c, K$ q& ?            Degraded data redundancy: 1 pg undersized
$ j! P, g, K) [  \% W
9 ~' `/ Z2 `, H! G: e8 G  services:
5 h. ?: u# ?1 U$ X( w, }3 C+ T    mon: 1 daemons, quorum controller (age 18m)
8 e4 x  z+ d% A6 Y) ?5 x7 k* t' L    mgr: controller.ifotuc(active, since 7m)
    osd: 6 osds: 6 up (since 12m), 6 in (since 12m); 1 remapped pgs
" U, v9 h4 Y* U9 @/ H0 d
  data:
. g5 v8 E$ _9 Z/ g2 k    pools:   1 pools, 1 pgs
6 \' [  u) {1 m5 x* v    objects: 0 objects, 0 B
" u3 h( y' Z0 |    usage:   6.0 GiB used, 114 GiB / 120 GiB avail
    pgs:     1 active+undersized+remapped
Degraded data redundancy: 1 pg undersized
3 T) d* k- L3 k8 E( h7 F

% l. x: \# c: [3 w
将调小size和min_size
ceph osd pool ls detail
pool 1 'device_health_metrics' replicated size 3 min_size 1 crush_rule 0 object_hash rjenkins pg_num 1 pgp_num 1 autoscale_mode on last_change 51 flags hashpspool stripe_width 0 pg_num_min 1 application mgr_devicehealth

ceph osd pool set device_health_metrics  size 2
) A6 w3 P% z# B4 D  H' s/ I
! J; O4 R6 Z. y+ f- i' hceph osd pool  set device_health_metrics  min_size 1
1 ~' d. _' I: N- t2 ^; I. L2 s5 ?set pool 1 min_size to 1
" k  Y/ Q& P7 J0 ^7 q$ ]
[root@controller ~]# ceph -s
4 K8 Q8 ?! x9 y1 C* Z# f7 I  cluster:
    id:     ed60a380-c9bd-11eb-ad78-000c29ad02de
    health: HEALTH_OK

5 V8 C; F  `5 s& S  services:
    mon: 1 daemons, quorum controller (age 17m)
$ l' g. |, B# c" d3 p7 D    mgr: controller.ifotuc(active, since 5m)
5 _3 O  h) S- K, S# ~( S* ]" b    osd: 6 osds: 6 up (since 11m), 6 in (since 11m); 1 remapped pgs
/ ^4 \1 B: p; d" ~& c4 W6 f3 Q+ Z
9 i+ y' e$ {) a$ m) x5 K# h0 V  data:
    pools:   1 pools, 1 pgs
8 h6 e, ^0 l, G: z% e; z    objects: 0 objects, 0 B
    usage:   6.0 GiB used, 114 GiB / 120 GiB avail
1 m! b/ O, Y3 j6 y7 o    pgs:     1 active+undersized+remapped
# D" w: \7 R, C4 _5 B

! H2 j$ N3 G; l. n6 L; e[root@controller ~]# ceph -s
  cluster:
    id:     ed60a380-c9bd-11eb-ad78-000c29ad02de
    health: HEALTH_OK

  services:
    mon: 1 daemons, quorum controller (age 21m)
4 M: D2 D6 A6 j7 l5 H9 C) O! W    mgr: controller.ifotuc(active, since 9m)
( b; l7 i$ z" _2 e/ y. E    osd: 6 osds: 6 up (since 15m), 6 in (since 15m); 1 remapped pgs
4 K8 C# G) J; W9 I: t8 k$ X3 x0 }
  data:
0 z9 \/ z9 @. H3 _$ N+ D# _5 u; G    pools:   1 pools, 1 pgs
    objects: 0 objects, 0 B
- Z+ |6 ?1 ?+ V* \' W2 w4 f4 c    usage:   6.0 GiB used, 114 GiB / 120 GiB avail
" V* m. S' A  @' M+ J    pgs:     1 active+clean+remapped
- `: u9 y, t! e+ b, _7 M
设置低于默认值就正常了。奇怪的问题。这是octopus版本的问题。
" ~3 b4 e& m1 w# Y/ ~device_health_metrics  是由cephadm安装后默认生成的。