ceph crash archive-all$ ?. b4 r% A5 ]* k/ N2 @2 O
[root@node03 ~]# ceph health detail$ `$ ~7 N- O& {7 t$ l( o j' N
HEALTH_WARN 1 daemons have recently crashed
3 r @1 y- i2 J1 M6 rRECENT_CRASH 1 daemons have recently crashed
Y/ C* q: t; B" K client.rgw.node03 crashed on host node03 at 2023-01-06 07:12:28.313044Z
- h: n: R% a ~6 I" C6 M; P4 ^) Q8 L% f7 C$ r
系统中所有的崩溃可以通过以下方式列出：
) a8 _. e# o' h2 [# M/ ^! o[root@node03 ~]# ceph crash ls; f: i- p; l! O e9 @- f
ID ENTITY NEW
4 f$ J+ X# V3 J+ V& w8 k2023-01-06_07:12:28.313044Z_df9f4fa8-d138-40f6-b8ec-c7c86299bb3e client.rgw.node03 * 9 l& Z0 w0 d s& o5 _
[root@node03 ~]# ceph crash ls-new
8 V) P* E: Q, uID ENTITY NEW
) T7 g1 X) g6 j5 l3 Z% A C! }2023-01-06_07:12:28.313044Z_df9f4fa8-d138-40f6-b8ec-c7c86299bb3e client.rgw.node03 *
, O8 G& Q/ l' E* |( u( ^5 N( N; I% l
有关特定崩溃的信息可以通过以下方式检查
7 o. P9 i3 c% E1 Q2 _[root@node03 ~]# ceph crash info 2023-01-06_07:12:28.313044Z_df9f4fa8-d138-40f6-b8ec-c7c86299bb3e
: X2 Z# f& T% Y. W" I{# N+ t. v5 C9 e# K. z
"os_version_id": "7.9", 5 O% w- { a- |& D7 q4 O' b ~
"utsname_release": "3.10.0-1160.el7.x86_64", ' ^% A, \ ^: ^) S3 S/ H
"os_name": "Red Hat Enterprise Linux Server", 6 {4 y) w$ y6 _4 `2 F
"entity_name": "client.rgw.node03", + Y: z+ } o6 Q
"timestamp": "2023-01-06 07:12:28.313044Z", ! ~" V+ F. k3 }& ^( F9 G4 H
"process_name": "radosgw",
2 x( ?: F0 s' {$ S "utsname_machine": "x86_64",
3 ?% J8 O3 c+ p3 ]! a6 I$ c "utsname_sysname": "Linux", 0 i+ e2 w8 z! m7 D- o' I2 V
"os_version": "7.9 (Maipo)",
: ?, H2 @* _" `' ~- \$ g+ P v6 t "os_id": "rhel", % h r) W7 n# v* y+ b0 P5 x
"utsname_version": "#1 SMP Tue Aug 18 14:50:17 EDT 2020", A: b; T4 F! {7 P, b, g H
"backtrace": [
6 ]% V3 S3 k" h* F: p( M( z "(()+0xf630) [0x7f5588ef4630]", ) J6 j# d- ]9 r* U g
"(()+0x84556) [0x7f5594e49556]", ' I* ]! [: J- t X e. O3 g4 g- g
"(()+0x86b0b) [0x7f5594e4bb0b]",
j- G9 @6 o6 ]% Y; B5 ~ "(RGWSI_Notify::unwatch(RGWSI_RADOS::Obj&, unsigned long)+0x2a) [0x5607b503643a]", & \$ ?, S# G0 y$ I2 k4 F# a/ g
"(RGWWatcher::C_ReinitWatch::finish(int)+0x34) [0x5607b503bb44]",
) d% Z/ S2 K( K& G; D5 H. i: K "(Context::complete(int)+0x9) [0x5607b4b59489]",
% K9 F& K% E4 a9 Y f "(Finisher::finisher_thread_entry()+0x16f) [0x7f558c0b8fff]", 3 _$ u# }$ g2 M; s: A/ m3 M
"(()+0x7ea5) [0x7f5588eecea5]",
% p! i$ V, F8 T X* Y+ M5 o9 y) i4 y "(clone()+0x6d) [0x7f55883f596d]"
2 G! P* b. M9 i ], N/ U* ~6 E: \1 S7 L" G3 r* _. ]
"utsname_hostname": "node03",
8 ^# X9 |, V# n: C1 { "crash_id": "2023-01-06_07:12:28.313044Z_df9f4fa8-d138-40f6-b8ec-c7c86299bb3e",
' U3 p" I3 D0 l! h0 r "ceph_version": "14.2.8-111.el7"
) a# k2 o+ U. f# Q- L5 ]0 O. c w: F}; R/ }2 W1 P5 J$ ` W+ y9 T! w, ~
1 [% u, M9 H' ` y
可以通过“存档”崩溃（可能是在管理员检查之后）来消除此警告，从而不会生成此警告
% e% n- a2 @# h% k0 w. A[root@node03 ~]# ceph crash archive 2023-01-06_07:12:28.313044Z_df9f4fa8-d138-40f6-b8ec-c7c86299bb3e + [' e) m- f8 i3 y
& r8 L. ^5 ?7 n- v. \5 m2 S
7 ^( G$ t1 t' T' z2 D
) Z" I6 D5 e- y- e% q& m, O通过ceph crash ls仍然可以看到已存档的崩溃，但不是ceph crash ls-new即可看到。 “recent”所指的时间段由选项mgr/crash/warn_recent_interval控制（默认值：两周）。 可以通过以下方式完全禁用这些警告： ceph config set mgr mgr/crash/warn_recent_interval 0 4 Z: E/ E) c6 u5 q
|
发表于 2023-1-9 15:00:11