找回密码
 注册
查看: 619|回复: 1

记一次模拟测试cephfs reset恢复状态 1 filesystem is degraded

[复制链接]

1

主题

0

回帖

12

积分

管理员

积分
12
QQ
发表于 2022-8-5 14:26:41 | 显示全部楼层 |阅读模式
[root@ceph-3 test]# rados ls -p metedata |grep 2.00000000 |xargs -i rados -p metedata rm {}
/ {. Z2 H* u  I( K  q: y  N8 u[root@ceph-3 test]# rados ls -p metedata |grep 2.00000000
# K) l3 n2 V2 V  X[root@ceph-3 test]# systemctl restart ceph-mds@2 M  z4 N0 N" p; i4 S
ceph-mds@                ceph-mds@ceph-3.service  
) a$ \5 A& M* M[root@ceph-3 test]# systemctl restart ceph-mds@
& ]+ Q  B  H2 x5 p4 Rceph-mds@                ceph-mds@ceph-3.service    }( r9 p. M% X: H7 C: ?& j7 {
[root@ceph-3 test]# systemctl restart ceph-mds@
1 }4 _4 Q' R$ i: p& i, nceph-mds@                ceph-mds@ceph-3.service  
4 X, ^  w! `7 S6 ^4 r[root@ceph-3 test]# systemctl restart ceph-mds@ceph-3.service
# A# R$ a( R; e/ H0 K6 I[root@ceph-3 test]# ceph fs status 9 {- S6 C, k5 f4 u( |; z2 A
                                         #Show the status of a CephFS filesystem
; w3 O4 x$ C, }- C- g2 t[root@ceph-3 test]# ceph fs status
4 B1 w+ _5 F: `8 M$ F                                         #Show the status of a CephFS filesystem
% C8 \) E0 P  f, v2 A/ i[root@ceph-3 test]# ceph fs status 6 p) Y* A% C9 W% F: ^
cephfs - 1 clients- T: `# ^' h+ L' Q
======
2 ]* p0 i; `7 x/ G+------+--------+--------+---------------+-------+-------+/ E, K- a* f5 o( d' Y
| Rank | State  |  MDS   |    Activity   |  dns  |  inos |
& [; p  L9 u0 P# f+------+--------+--------+---------------+-------+-------+
) I; M9 o; n% j3 h|  0   | active | ceph-1 | Reqs:    0 /s |   11  |   13  |9 f* d- J  x6 P# I  {2 M3 B
+------+--------+--------+---------------+-------+-------+
3 ?& ?8 R( `' O) I+------------+----------+-------+-------+
7 V) [, A4 A  c: s|    Pool    |   type   |  used | avail |5 {5 V8 f( A$ J
+------------+----------+-------+-------+
# `$ a. R. E+ f2 p5 c. S4 b5 o  Y6 R|  metedata  | metadata | 5911  |  277G |; K' U. O7 r" T! Y7 y( V9 [
| cephfsdata |   data   | 44.5k |  277G |+ L# [# D/ d2 r, {, y, M% z
+------------+----------+-------+-------+
* U  o# a8 Z" {, s
- R9 \0 E! A) d7 o' Q& s- j+-------------+4 n- F0 z, K  L
| Standby MDS |
3 ^+ U8 ^/ }& j& ]: ~  `+-------------+
* n( w7 a6 l7 e, [|    ceph-2   |8 s/ I- _) e- w* C; W0 p& T5 w( ~
|    ceph-3   |
/ ?: r! N8 {* i4 \# G  f+-------------+: }/ U1 d, Q* w6 @) [2 Q6 n
MDS version: ceph version 12.2.8-52.el7 (3af3ca15b68572a357593c261f95038d02f46201) luminous (stable)
' A# H5 w, S9 y3 V7 D- ?0 Y! _[root@ceph-3 test]# ceph fs status ) ]1 z5 y. S- B& J& d; C0 L
cephfs - 1 clients- A+ M* A6 I" z- T$ A/ [/ ^
======
, l- U; m: g+ g, l5 R8 X. U+------+--------+--------+----------+-------+-------+2 p0 t5 C, l; k6 v. I
| Rank | State  |  MDS   | Activity |  dns  |  inos |- b, O+ M  r: S: L4 h# H
+------+--------+--------+----------+-------+-------+5 q# D. j( a% x, B% q, Q& p4 h/ p
|  0   | replay | ceph-3 |          |    0  |    0  |
7 D4 B. ]! i, j+ q  G# W1 D+------+--------+--------+----------+-------+-------+' K% K' U9 H, o. m
+------------+----------+-------+-------+% Z1 J$ o# h9 c; z9 R" n0 a2 j
|    Pool    |   type   |  used | avail |
# v# K9 ^; f1 R2 e$ |4 u3 H! `. O+------------+----------+-------+-------+
5 D+ ~* l9 q; |9 ]+ O9 n+ q" `|  metedata  | metadata | 5911  |  277G |
' ]) `9 i0 P8 W| cephfsdata |   data   | 44.5k |  277G |
( r3 L' f" m$ T; X+------------+----------+-------+-------+
7 K) d( T( X# r- F. ^: s1 z3 x- s- r' Z1 x
+-------------+
; E/ q  W. M6 H+ D| Standby MDS |: l7 M# F5 x! L1 x, ^
+-------------+! B) }; b4 \; F3 [
|    ceph-2   |8 c, r# P& T( l
+-------------+
. m! P4 L/ \4 P0 h; L' wMDS version: ceph version 12.2.8-52.el7 (3af3ca15b68572a357593c261f95038d02f46201) luminous (stable)
  U) i! _0 |# D. _' b6 l% c[root@ceph-3 test]# ceph fs status 5 V. }6 t1 [& p* u' N; x
cephfs - 1 clients
- e" `: i* S9 ?) f! F- @! b======7 [  \- H, [2 ?  j
+------+-----------+--------+----------+-------+-------+
1 L2 f" P! J/ y) x% e8 H7 S| Rank |   State   |  MDS   | Activity |  dns  |  inos |4 `8 b/ A* x2 S, Y- G
+------+-----------+--------+----------+-------+-------+6 d, H. y6 s( R: J3 ^
|  0   | reconnect | ceph-3 |          |    0  |    0  |
0 s' q9 }  x0 R: O( s/ n) k+------+-----------+--------+----------+-------+-------+
7 Z0 \4 ]5 u+ p+ j% E0 I4 V; P+------------+----------+-------+-------+
! P1 q4 q) n: O+ Z|    Pool    |   type   |  used | avail |/ @- J5 u, K  g& l9 ]5 [0 Y1 ]" ?
+------------+----------+-------+-------+/ x) O8 f, y; E* w' L* n
|  metedata  | metadata | 5911  |  277G |6 l) e2 u" b9 L* g
| cephfsdata |   data   | 44.5k |  277G |  p) I6 S' i; Q8 E$ b3 ^
+------------+----------+-------+-------+
2 E. \8 c# ?+ f: B5 L: g3 Y, h8 I( S; W
+-------------+" s* a3 q+ L6 S
| Standby MDS |
, P- Y  V  }1 _1 K* h% L+-------------+! X( _! }$ a: q) f( S
+-------------+
& z2 V) _1 n2 C" EMDS version: ceph version 12.2.8-52.el7 (3af3ca15b68572a357593c261f95038d02f46201) luminous (stable)4 f# X5 j4 k' A% i5 E# l6 V: G
[root@ceph-3 test]# ceph fs status
$ F: N8 `3 v5 t% q. s. acephfs - 1 clients* h9 S1 N$ l, n* j
======
( t& k0 C! f  _* Q& q' b$ W+------+-----------+--------+----------+-------+-------+0 i  e; K4 ?" M8 @- n- e
| Rank |   State   |  MDS   | Activity |  dns  |  inos |
( P3 A6 q# @6 ~2 `( n+------+-----------+--------+----------+-------+-------+) ?( e0 a" T/ t5 M* k& x5 ]
|  0   | reconnect | ceph-3 |          |    0  |    0  |0 Z" C+ U3 e, t
+------+-----------+--------+----------+-------+-------+  w  _$ G: O5 [6 A& P6 o) k1 u" e
+------------+----------+-------+-------+  ]% x) e, `- L% J, p! U1 t! {0 u9 v( }
|    Pool    |   type   |  used | avail |
" H; G$ ]5 ]$ v: [; i7 a+------------+----------+-------+-------+
  p7 [  z. x) }' M" j1 k+ [|  metedata  | metadata | 5911  |  277G |
$ @+ O3 Q7 I. h( M' e  J$ I/ _| cephfsdata |   data   | 44.5k |  277G |# W7 j0 h/ X0 T* g. W
+------------+----------+-------+-------+
% g8 g3 L5 o, U' k( G" f6 @7 K$ H
+-------------+
0 F2 X: y1 k0 I; w6 p3 n# L5 K! V| Standby MDS |$ Y% S* N5 B) S6 C1 i# w" L
+-------------+
* H, z8 O3 e3 r0 V+-------------+
* \' D2 A" j7 o. a) d. GMDS version: ceph version 12.2.8-52.el7 (3af3ca15b68572a357593c261f95038d02f46201) luminous (stable)
. \2 d9 {( P. V) n) I[root@ceph-3 test]# ceph fs status ; x* V, A/ ^  {" Q( B# x
cephfs - 1 clients
# O& R1 m% P- X* a======
* ~0 x/ |3 a% G; _$ f4 f6 I  ]7 Z; [+------+--------+--------+----------+-------+-------+
% S: D$ j9 @$ Q* F7 A| Rank | State  |  MDS   | Activity |  dns  |  inos |
8 `/ c+ c* o8 @8 Q& I  f0 N+------+--------+--------+----------+-------+-------+7 H5 O8 D* y2 P8 e" J1 `$ f+ s3 T
|  0   | rejoin | ceph-3 |          |    1  |    3  |% w2 X5 x+ d9 y! o3 q% H
+------+--------+--------+----------+-------+-------+7 k) j& l0 j. z* V/ E& H5 l
+------------+----------+-------+-------+
7 z* l: g+ ?0 S+ v2 H4 B|    Pool    |   type   |  used | avail |- ~4 K3 p4 Q8 o3 ^& a( A! O
+------------+----------+-------+-------+1 ^" F1 I. F! g5 @( h
|  metedata  | metadata | 5911  |  277G |% Y' @) m/ Q3 K
| cephfsdata |   data   | 44.5k |  277G |
; W! ^1 I; k) p  T- d" Z+------------+----------+-------+-------+
4 T1 ]: N% C: Q$ i2 ^& t7 U4 Q" C9 ~! h1 b3 }9 x
+-------------+
& [( n7 j* w; g, k  W7 A| Standby MDS |1 H; M1 _8 P& q! c5 f5 D" d2 C
+-------------+
! Z2 |  r5 Z1 [( t2 l* P; E|    ceph-1   |
# R0 Z0 g" w* o' K& e+-------------+  j9 q7 q; Z" S/ f3 g
MDS version: ceph version 12.2.8-52.el7 (3af3ca15b68572a357593c261f95038d02f46201) luminous (stable)
/ U& `+ h1 g2 w5 R; L* \! B! O[root@ceph-3 test]# ceph fs status " q/ _" D0 i; s* ^: N
cephfs - 1 clients% y8 C. s7 Z8 W+ T- L. e
======: i5 m6 E( U% L( E, c4 r, x9 y8 o
+------+--------+--------+----------+-------+-------+
1 J$ `0 {3 i" L& Y6 g* c| Rank | State  |  MDS   | Activity |  dns  |  inos |
  I# y& [: J. V: d) S$ f# s! ^7 d& d% [+------+--------+--------+----------+-------+-------+
- q! F3 @8 g0 p4 Z# u|  0   | rejoin | ceph-3 |          |    1  |    3  |
$ T) P& Z; k7 U5 R7 q, v) f: j4 g6 a+------+--------+--------+----------+-------+-------+
- e* H! j6 }. y/ r5 @( Z/ N+------------+----------+-------+-------+5 Q) V9 y+ g& e2 F2 ~, r, ~  j
|    Pool    |   type   |  used | avail |: @1 r8 E+ B* C; r$ U. c0 b+ i
+------------+----------+-------+-------+
! a* C: X- v5 t" b  Z|  metedata  | metadata | 5911  |  277G |0 F; G8 |; f+ J' P/ [
| cephfsdata |   data   | 44.5k |  277G |; D3 O3 x; |3 N6 j
+------------+----------+-------+-------+
1 f! @& y9 n5 R) _0 `! q
) x, C& v' D1 c2 d+-------------+: f% t" {% H. u: ?+ I
| Standby MDS |* i! `9 A" c* ~  b
+-------------+6 a8 e) Q1 J2 D: l
|    ceph-1   |
0 H, H% r  f% P" K' ]3 F+-------------+( @3 J$ W8 N! `3 M
MDS version: ceph version 12.2.8-52.el7 (3af3ca15b68572a357593c261f95038d02f46201) luminous (stable)* y/ s% M5 `6 Y
[root@ceph-3 test]# ceph fs status
' ?' p: g, L: h; j: q/ n0 ^cephfs - 1 clients
+ [+ E; e" Q5 X% Q+ v8 _======& y/ I7 a5 A( r) Y$ R, r" w
+------+--------+--------+----------+-------+-------+& {6 ~' O% z' S3 `0 ]1 d
| Rank | State  |  MDS   | Activity |  dns  |  inos |' F' J8 a, D$ y+ h- o
+------+--------+--------+----------+-------+-------+# U' U- C5 b* F" {: O) N
|  0   | rejoin | ceph-3 |          |    1  |    3  |
. N' n1 b  C; w# j% k' I1 l+------+--------+--------+----------+-------+-------+
3 D0 ]& L2 p! G' |; I: u+------------+----------+-------+-------+8 G: i/ g* {& E3 d; |3 a
|    Pool    |   type   |  used | avail |, Y/ ^2 ?7 n* h* R% C# d6 t3 T
+------------+----------+-------+-------+( Y. C) J& p4 I$ N
|  metedata  | metadata | 5911  |  277G |
' c5 n9 M; S  [; l% c| cephfsdata |   data   | 44.5k |  277G |. m  s' f5 r( J7 r; b2 }9 {
+------------+----------+-------+-------+" ?6 X" n( N" L9 a2 Q9 Q1 t4 s

4 L' j8 h  W# `+ ?+-------------+
: R! s8 L. _$ m; ~( W) H| Standby MDS |
$ X, x0 _+ j  A- q& V* b' Q+-------------+
( l4 p/ F+ T3 w|    ceph-1   |9 j8 d9 `7 h2 a5 R( L; b* a
+-------------+
; n0 G6 D' {0 S; cMDS version: ceph version 12.2.8-52.el7 (3af3ca15b68572a357593c261f95038d02f46201) luminous (stable)7 i' ]' n# ~9 v4 z6 ^8 F! Y
[root@ceph-3 test]# ceph fs status ! k1 t3 J$ a+ x4 C, y
cephfs - 1 clients* ?3 P6 I) A" K( X9 O6 U) ?+ e
======
( l$ q! ?3 ?/ S4 e- P+------+--------+--------+---------------+-------+-------+
5 o9 N" ]! l2 F* r| Rank | State  |  MDS   |    Activity   |  dns  |  inos |
' |( v' ~: q5 Z" [8 L+------+--------+--------+---------------+-------+-------+  Z8 z4 X' `. t
|  0   | active | ceph-3 | Reqs:    0 /s |    1  |    3  |
  T4 E" Q  X# n+------+--------+--------+---------------+-------+-------+
& q! u% f: |; ^# m: G+------------+----------+-------+-------+
7 [/ |0 O7 a, T. G* [+ q) }|    Pool    |   type   |  used | avail |
# V3 w- E% @7 F6 `* z+------------+----------+-------+-------+; G2 c6 |* p1 F1 @1 _) s) `8 N
|  metedata  | metadata | 5911  |  277G |; V4 S: E. a6 a% J
| cephfsdata |   data   | 44.5k |  277G |
$ S- f1 X$ y/ n: ]2 ~+------------+----------+-------+-------+8 u# s' ^# Y% q1 n6 x4 g! o
/ s( O; h. ^# c
+-------------+1 Z. H( w$ J* O
| Standby MDS |
( |2 _3 I. X, X% \4 {+-------------+
. W; P# w" w6 f6 ||    ceph-2   |. k1 @8 @5 R, X4 @1 v) B
|    ceph-1   |
! i& X( f; J- T; L" D- B+-------------+6 [  G0 O; N; Y# w" {2 Q, c# S
MDS version: ceph version 12.2.8-52.el7 (3af3ca15b68572a357593c261f95038d02f46201) luminous (stable)
& d1 d7 M" z9 ~[root@ceph-3 test]# rados ls -p metedata |grep 2.000000008 ~) `% O8 A0 _5 p
[root@ceph-3 test]# rados ls -p metedata |grep 2.00000000. \5 D+ n( D- S+ M* w' i
[root@ceph-3 test]# ls
4 w3 ^! G# p2 u' S2 i# T3 f10000000000.00000000  recoverfile.png
5 Q' y) E- z& m* Z" ~( R  R[root@ceph-3 test]#
' l& e, P/ X0 j  H[root@ceph-3 test]#
7 j) P" l# J. g/ V2 t[root@ceph-3 test]# cd $ m' X$ X! h- k, [( f
[root@ceph-3 ~]# rados ls -p metedata
' ~5 x7 u7 w6 C; B601.00000000
2 u8 e% a5 M7 `. t' n- p600.00000000& c$ v' U. o" N4 U
603.000000007 n% ]2 f* q1 o. [* G3 k) b- t8 ~/ Y( N
1.00000000.inode
6 w+ P. s3 O9 T. [200.00000000
! Y/ r: q- [# c: l200.000000015 {. X& \- X" T
606.000000002 g( [3 ?% V8 @* J, q4 Z9 Z
607.00000000
, b. Z7 I: u8 a4 o2 d, B/ v) G1 h608.00000000, _# T+ ~" y8 g1 i: }8 A
604.00000000
/ I& X, ^: z6 p500.00000000
8 t( {. Q: Z1 E% b2 W$ Vmds_snaptable
# k- m/ ]2 h0 q" p, G605.000000007 B, K& u: n2 e) `! ~( D
mds0_inotable
% ?) B1 h  K: u100.00000000
1 Q8 |# s4 K2 S* T: K; Xmds0_sessionmap. `3 k6 O2 Z4 k6 ^& c' @
609.00000000
+ W+ c, @0 [0 c$ @$ ?+ ?- S400.000000006 f. Q. i- _+ C
100.00000000.inode
) U2 H( B4 u+ \; b4 \4 U$ Z1.00000000
4 B- B; `) Y$ p2 c" S" @9 M1 {. Q[root@ceph-3 ~]# rados ls -p metedata |grep 602
7 c  ^8 X& ?2 X; N[root@ceph-3 ~]# cd /mnt/^C* {( y5 K- {0 P& Y& _# g
[root@ceph-3 ~]# ceph -s
4 l# ]2 Z6 L3 V; i6 ?  cluster:
0 Y1 q2 f& O5 m9 }3 w9 l1 S1 x    id:     57bf4711-2218-46af-99d6-9c68ae230ce17 e) {! }0 C0 K2 v3 D5 ~
    health: HEALTH_ERR7 ~" Q" W- \; X( ^  J" h! ?
            1 filesystem is degraded
% m# S, ]+ G/ c' g4 @* a" Y) ?" z            1 mds daemon damaged) x% s1 K, t" D

, \) e- A: B% V1 ]5 O: `  services:
  L( V1 h9 g9 `% K3 A    mon: 3 daemons, quorum ceph-2,ceph-1,ceph-3
4 U8 G/ S% {' u6 A/ k; e* s6 \3 O    mgr: ceph-2(active), standbys: ceph-1, ceph-3
2 g$ O2 z+ C) {& h7 V* I    mds: cephfs-0/1/1 up , 3 up:standby, 1 damaged
; x4 V1 ~) c5 a/ e' m    osd: 6 osds: 6 up, 6 in9 d$ i1 q9 p9 |1 }! P9 O! T" p

3 f5 o$ Y; t0 M3 R% i4 K  data:* ~% Z8 [+ T$ }3 s9 Z) ?, j7 A! \
    pools:   2 pools, 256 pgs9 G& f7 i5 G1 _% a% t% T5 Y3 V
    objects: 21 objects, 50.4KiB1 i; E! I( n3 L  i# s1 ~
    usage:   738MiB used, 584GiB / 585GiB avail
( D+ H6 g0 B5 F5 B    pgs:     256 active+clean8 U& K: r" e/ s. R

% c6 }' P4 p& `8 @[root@ceph-3 ~]# ls
4 N1 `4 K8 x. |- z, M: k/ F# f) o7 Brecovery_cephfs.py+ M  h% S$ v/ a7 l$ o
[root@ceph-3 ~]# ls
- F$ w( F" w0 }  I; b8 S: Vrecovery_cephfs.py
  C) D, g/ k, a/ w/ Z6 x[root@ceph-3 ~]# 6 d8 W/ o2 v2 L4 H( b3 a; l( T
[root@ceph-3 ~]# ls
6 W4 j; X- E* e1 I" krecovery_cephfs.py; {4 O& B! e; k4 B3 I& \% n" B
[root@ceph-3 ~]# df -TH
- Y5 K# h1 T* o0 u+ i2 \$ G% l$ c4 ^- x4 u& o! e
: Y8 L  [$ _7 S2 z2 g3 T2 `# g
7 G- B/ a# K+ d
^C" _& m* f6 d7 @2 C& A
[root@ceph-3 ~]# ls
2 t6 G- A: c/ r9 u; ~1 yrecovery_cephfs.py0 s) ]* x; R) K/ l$ I
[root@ceph-3 ~]# ls
; u1 o# W% c9 q3 ^' g7 H9 zrecovery_cephfs.py
0 \* i; L9 ~" b$ X$ t! k[root@ceph-3 ~]# ls
' D+ y. z, O. a9 K! U) r/ D7 ]1 ?recovery_cephfs.py, f8 o# J( T7 l- O
[root@ceph-3 ~]# ceph -s" w6 w. ]& @, X; M
  cluster:
, }" e) p( e; Y; }# t1 O7 D    id:     57bf4711-2218-46af-99d6-9c68ae230ce1
8 Q1 K7 s$ Z# ]# p9 s8 i    health: HEALTH_ERR
: c% x: }0 s# W" c* D            1 filesystem is degraded. x3 z, D& i' i* ]. t- q7 M8 h
            1 mds daemon damaged. e$ M5 v$ K$ G

! h4 u' |+ t4 d- @2 ?* R1 ]3 U' p  services:
2 n0 ^, v0 a4 p: Y6 G    mon: 3 daemons, quorum ceph-2,ceph-1,ceph-3
( }# H; A4 ~' V+ x% u8 x& u    mgr: ceph-2(active), standbys: ceph-1, ceph-31 {( w" I1 R7 s. v$ _
    mds: cephfs-0/1/1 up , 3 up:standby, 1 damaged5 ^  I! Z8 s! C% Y4 v( P/ U
    osd: 6 osds: 6 up, 6 in
3 F0 h4 |$ ~$ h' P3 b: K* E! W7 u3 v# Z2 b9 Q7 q: Z) [" K
  data:
( Z8 q4 g( W( k# I3 V    pools:   2 pools, 256 pgs
6 O; j2 U. S. X+ g6 n    objects: 21 objects, 50.4KiB
! n0 m4 X# t0 C7 b    usage:   738MiB used, 584GiB / 585GiB avail! i: [2 v3 X& Y- d
    pgs:     256 active+clean9 ~) O5 K! w' y

  I+ ]6 G3 {1 P; D[root@ceph-3 ~]# ceph fs status s) ]  a- h( s( j8 W! I
3 e/ z) B" v4 z9 K0 g8 T
+-------------+
0 [$ Y" i! r/ V: Q| Standby MDS |
/ B/ e5 C3 z9 r' T5 t+-------------+
+ N- N4 O- h( @/ i  Z|    ceph-2   |
: v5 g& z' A$ w6 c. `8 J" A& y|    ceph-1   |
1 l+ `6 ]2 M2 M4 u0 N|    ceph-3   |
( s2 B7 `$ H; T6 m4 ~+-------------+
1 S; L3 ]2 D# ^) AMDS version: ceph version 12.2.8-52.el7 (3af3ca15b68572a357593c261f95038d02f46201) luminous (stable)
( N+ R0 i: Y: o' u5 R[root@ceph-3 ~]# ceph fs status $ c0 U) t7 H: s) k. `
cephfs - 0 clients. s( w5 ]* |. k( Y
======: o& _4 I' x" w- k
+------+--------+-----+----------+-----+------+
5 k5 e5 f  X/ V" B3 l6 r: ], P| Rank | State  | MDS | Activity | dns | inos |
! E4 u8 D; R1 V/ N5 [; A" l( J+------+--------+-----+----------+-----+------+$ `( ?* {' }& b+ D' A# J
|  0   | failed |     |          |     |      |! w/ `+ I& ^7 z9 `, J
+------+--------+-----+----------+-----+------++ \8 J# i7 d2 G" f& G8 }7 K3 }
+------------+----------+-------+-------+/ o) x7 t; }5 C
|    Pool    |   type   |  used | avail |& I7 @% f$ v! W: W+ f
+------------+----------+-------+-------+
0 k8 y2 @5 J. z5 \  ?# N3 `" m|  metedata  | metadata | 6023  |  277G |
1 v9 j: s/ \/ ^9 R- j| cephfsdata |   data   | 44.5k |  277G |
0 r* ?) v! ?: i1 l9 ~+------------+----------+-------+-------+& v! d  b6 r' }. T. t+ W% U

& r. P3 B* j& T& y# _" q& P. G: \+-------------+
4 c% O% y& D4 L) A| Standby MDS |: J/ D) w4 ]8 G' U& e) ]) O0 F4 s
+-------------+
6 E  v" G; _6 `; }|    ceph-2   |
6 e" s6 g1 U4 F+ r3 ?  q|    ceph-1   |
6 C9 f) J" O5 M9 {" h|    ceph-3   |0 d. c$ y+ c- q6 j5 K) H
+-------------+( I. g0 d; k% }7 \
MDS version: ceph version 12.2.8-52.el7 (3af3ca15b68572a357593c261f95038d02f46201) luminous (stable). S9 T  j  ^+ v! [# f7 I
[root@ceph-3 ~]# ceph fs status
1 z8 t% N( `1 N" r6 D0 L  Scephfs - 0 clients. A  N$ Q+ g/ \' }; \
======1 j" ]5 T9 e5 T8 X- W7 E' N5 t  Y
+------+--------+-----+----------+-----+------+6 {) X2 L7 U" [9 b& _2 d" c, ?% w$ Y6 x7 ?  ?
| Rank | State  | MDS | Activity | dns | inos |" O8 p- j1 k( R$ O$ }$ z: P8 l
+------+--------+-----+----------+-----+------+
% {# u9 g2 F' c# U  ]|  0   | failed |     |          |     |      |( n$ @7 }* n/ M/ `( `6 e
+------+--------+-----+----------+-----+------+
7 A# t2 C3 p( p, T+------------+----------+-------+-------+1 l; P/ A5 @: @: n5 Z* c
|    Pool    |   type   |  used | avail |% D4 [$ `. G. Z8 _# V5 H
+------------+----------+-------+-------+7 n. W7 j. j! M. f
|  metedata  | metadata | 6023  |  277G |
4 e# {8 \  A, ^; V4 Z% K- Q| cephfsdata |   data   | 44.5k |  277G |
# f. ]0 [8 y7 I  ^4 _5 Q' k+------------+----------+-------+-------+
3 C. {3 v% b) h5 f+ K/ b' A! }7 j; n; `% l$ w7 M/ R
+-------------+
* w2 ?7 m( E2 @+ T. t| Standby MDS |3 N' [; y! w. a& y& G  K
+-------------+
( y3 L( J$ {2 k$ E- g/ L|    ceph-2   |9 Z6 }" w' {0 d2 y+ R
|    ceph-1   |& q2 Z7 ?5 D4 i' R/ {7 l, m  F9 y
|    ceph-3   |" f# A. t- t! p, a5 n! l  P+ R
+-------------+! x0 h0 c( j) ?3 Z3 |' }% z
MDS version: ceph version 12.2.8-52.el7 (3af3ca15b68572a357593c261f95038d02f46201) luminous (stable)6 }1 C; F# Q7 {! X' u
[root@ceph-3 ~]# df -TH
" [, a) R# R6 L^C
% y5 ?" D/ X  K! |, Y[root@ceph-3 ~]# ls
: p3 k0 f8 P( _! r1 @recovery_cephfs.py
0 I0 A0 G- @9 e- l+ h[root@ceph-3 ~]# . F8 J( `( q, w! O3 Y, [( A
! b& L# ^# n( \
[root@ceph-3 ~]# ceph fs ls; h2 h5 {( J! Q& z3 D+ n+ A3 V2 d
name: cephfs, metadata pool: metedata, data pools: [cephfsdata ]
8 S4 t& l) d' ]. s[root@ceph-3 ~]# ceph fs reset cephfs -
, \$ K9 F7 [/ F, n2 t--admin-daemon     -f                 -n                 -v                 --watch-info. K  S( b1 t: b! D, A% t
--admin-socket     --format           --name             --verbose          --watch-sec( i" |9 U' g8 Q: O% A
-c                 -h                 -o                 --version          --watch-warn& J- a  @1 x& g/ H" `2 y+ \
--cluster          --help             --out-file         -w                 
8 E1 Q# F* ]: u' f$ U2 \& O--concise          -i                 -s                 --watch            
  Z" T7 e/ H* K. e--conf             --id               --status           --watch-debug      
! C) i& w# _/ S+ p. a--connect-timeout  --in-file          --user             --watch-error      - G: Q: N4 k3 e" n# J5 ]
[root@ceph-3 ~]# ceph fs reset cephfs -
& a* n. v$ n: _6 X! t2 `! ^--admin-daemon     -f                 -n                 -v                 --watch-info$ b3 m* h3 i1 |) N4 d
--admin-socket     --format           --name             --verbose          --watch-sec
* X# |, P) Y7 u7 J! ~6 X; J-c                 -h                 -o                 --version          --watch-warn+ }2 W. m0 `. [9 `; u7 F
--cluster          --help             --out-file         -w                 
3 F$ G: P( w- j5 A2 d--concise          -i                 -s                 --watch            ' {, h" m2 G5 p. n" X" W1 J5 i( l, _
--conf             --id               --status           --watch-debug      ( @" J6 n; `4 ]$ R% a6 u
--connect-timeout  --in-file          --user             --watch-error      . y4 v& a; G) ^0 k5 T" ]5 N! f+ S6 w% t
[root@ceph-3 ~]# ceph fs reset cephfs --yes-i-really-mean-it , h0 J: ?* S3 W. O% U
                                                    #disaster recovery only: reset to a single-MDS map
% @+ ~: A/ b7 Q& I$ j: {[root@ceph-3 ~]# ceph fs reset cephfs --yes-i-really-mean-it
+ M3 q- w: r2 a8 V; J( i$ c$ n[root@ceph-3 ~]# ceph fs ls
9 a+ J, x5 w( Fname: cephfs, metadata pool: metedata, data pools: [cephfsdata ]  ~: Y( U2 |4 n4 A* F& z  C  u; D# z- {
[root@ceph-3 ~]# ceph fs status
8 t+ R9 G6 r4 k- K. Rcephfs - 1 clients4 j/ ~$ h1 [' b7 |7 y% `% {
======4 E' y+ E5 V+ P" y# d
+------+--------+--------+----------+-------+-------+6 R! W# X" G+ }0 }  G
| Rank | State  |  MDS   | Activity |  dns  |  inos |# y* P% r# Z" R7 J$ y
+------+--------+--------+----------+-------+-------+2 b4 e7 o0 p; t; G/ C
|  0   | rejoin | ceph-3 |          |    1  |    3  |6 _9 k- j1 f& W* U% a  X3 V
+------+--------+--------+----------+-------+-------+
' J/ T; s, C- z8 l+------------+----------+-------+-------+# _9 W8 V, c0 }
|    Pool    |   type   |  used | avail |( w& j7 |  J0 L5 q# E# n. I% a( Q
+------------+----------+-------+-------+
7 W4 z* M8 b' v* D2 i|  metedata  | metadata | 6023  |  277G |9 t4 Y, J+ {' x. N
| cephfsdata |   data   | 44.5k |  277G |& \- p/ q& I. F9 L2 Q. k
+------------+----------+-------+-------+
- J' d0 s; D5 l# d; N% @% i' L+ u
+-------------+
- i7 \3 \" ~* z' A0 w| Standby MDS |
* o' B0 y) E0 C4 i, W+-------------+
$ ^1 P2 _' i$ B0 O|    ceph-2   |
' N/ S5 \  K7 ^, h) x$ h+ n) _|    ceph-1   |! l3 h% L) t) J2 V  Y9 B3 F9 r2 q2 o( i
+-------------+
4 D1 j7 P  P$ V5 [MDS version: ceph version 12.2.8-52.el7 (3af3ca15b68572a357593c261f95038d02f46201) luminous (stable)
4 U& a5 v9 R% O4 O* M[root@ceph-3 ~]# ceph fs status
( P/ p# \/ z" V; X3 m. A6 [cephfs - 1 clients, @5 I; g# l; I) I: B
======4 ~* }' z4 n# F. z
+------+--------+--------+---------------+-------+-------+
- h' }7 A; J- k0 V| Rank | State  |  MDS   |    Activity   |  dns  |  inos |
* Z% ?3 O; n4 [% g4 B( y+------+--------+--------+---------------+-------+-------+
7 D; o( N) n4 s  V' }2 n5 E$ k( W|  0   | active | ceph-3 | Reqs:    0 /s |    1  |    3  |
% |- U. G# c- o4 j# N+------+--------+--------+---------------+-------+-------+
& {8 D5 L$ _; s. @+------------+----------+-------+-------+" r8 D1 _1 O9 \7 N) {; T, V
|    Pool    |   type   |  used | avail |
2 i$ Q4 C0 D+ c9 V2 J7 Z+------------+----------+-------+-------+
* K$ v' m% \  q9 ~4 z* u$ k|  metedata  | metadata | 6135  |  277G |+ s5 P. x' {3 G$ A9 |# A+ D
| cephfsdata |   data   | 44.5k |  277G |
* B6 H0 |% N) U" T* @/ [+------------+----------+-------+-------+' g& _; I! f: Y; W% J
& v- s* f% i! U( E% R6 H7 O6 N1 `+ n
+-------------+
& G2 h, d) T! }# ?  o9 p| Standby MDS |
. `* }" G! U6 I) k+-------------+
9 v7 F9 x2 z* c$ t|    ceph-2   |
- o* G! l9 F) ]$ I+ G& Y|    ceph-1   |, I" U9 ]) l# e8 `
+-------------+3 W  r' P; T+ B! Q
MDS version: ceph version 12.2.8-52.el7 (3af3ca15b68572a357593c261f95038d02f46201) luminous (stable), V, h4 N5 f- A4 Y
[root@ceph-3 ~]# " B! j( X. c: G  H# F" b
6 Z) W: s1 Y, O4 a  n0 T

! z. }* _$ m' S2 s6 r/ ]% a0 [

1

主题

0

回帖

12

积分

管理员

积分
12
QQ
 楼主| 发表于 2022-8-11 09:25:06 | 显示全部楼层
解决:
& s: l7 f# p. _  p- ]2 }针对原因一:6 R! d. Y. l4 U3 V
, p2 A1 y5 I- Z1 B
此次情况比较特殊,ceph health detail命令竟然看不到stuck的osd,解决如下:0 [& t$ T3 N4 w8 N) t: Q" G9 `
把 使用率比较高的osd降权重,手动触发osd的数据迁移。
7 N; `+ z+ c6 ?: e& z/ M重新ceph health detail命令,此时会发现stuck的osd,重启相应的osd,mds状态恢复。3 [- K; X6 ^& C( \6 C
针对原因二(收效甚微):
) m$ T$ o- d1 t7 |$ k5 J, A
1 ]$ x. i9 {) [! a; \1 f/ I! @调小mds mds_cache_memory_limit到40G 以下
3 W7 J' Q3 U& ?9 Y# ceph tell mds.\* injectargs '--mds_cache_memory_limit=40000000000'
3 w+ T& l" E2 R+ @& X$ y: p在/etc/ceph/ceph.conf修改,传到所有ceph节点
4 j: a) h) Y- I# p2 m' ~针对原因三(主要原因):/ s/ Y9 J0 s  b: m/ k& |
# r, I4 B+ Q* a! |
方法一 mds多活热备调低敏感度,有两个操作:  c" k, d9 \- {7 x! g# W. f

8 v; R0 l4 X% h8 E8 N  e: i调高mds_beacon_grace=300,调高超时时间。6 E9 W* ?) j/ Z( i/ {' r

' ?0 R! D( a; l# U, O' N& F; v  写到/etc/ceph/ceph.conf里,传到所有mds节点,然后重启mds9 |7 i2 j7 e: q) ]

1 e6 q) ?% Q8 X- B: Gmds采用热备模式,替换原来冷备模式,加快mds加入速度。2 J/ s+ F, M/ r. x+ q2 d$ }
7 W( w" s9 u3 q0 t
  ceph fs set <fs name> allow_standby_replay true  U% l0 `0 t( C/ T

7 g" \$ Z: D2 T0 s# ^) J方法二 mds多活热备静态目录方法,即给目录划分不同的mds  l* e; p3 ]* d
' w1 l- n1 h4 w9 }5 d" n7 b5 E9 l
命令:
6 ~, u! M5 E# }9 [setfattr -n ceph.dir.pin -v 2 path/to/dir+ W* |7 D7 S! C  @0 z
mkdir -p a/b$ Z8 {. l  K8 A
# “a” and “a/b” both start without an export pin set# ~: Z. h1 N  t5 ]
setfattr -n ceph.dir.pin -v 1 a/
! Y. k4 X5 S9 H# \: d2 I- l# a and b are now pinned to rank 1: g9 V8 V) V- [/ j
setfattr -n ceph.dir.pin -v 0 a/b
1 E( [( {7 c$ S2 |# a/b is now pinned to rank 0 and a/ and the rest of its children are still pinned to rank 1
% {. k& q) ?* e% `. W! Y) P# x' L+ b  ]+ q; s) U2 A
( S. j8 t5 F/ w$ @$ M! G8 o
您需要登录后才可以回帖 登录 | 注册

本版积分规则

返回首页|Archiver|手机版|小黑屋|易陆发现技术论坛 ( 蜀ICP备2026014127号-1 )

GMT+8, 2026-6-12 00:51 , Processed in 0.019005 second(s), 22 queries .

Powered by Discuz! X5.0

© 2001-2026 Discuz! Team.

快速回复 返回顶部 返回列表