osd 无法启动
问题描述
osd 状态 down
[root@node-1 ~]# ceph osd tree
ID WEIGHT TYPE NAME UP/DOWN REWEIGHT PRIMARY-AFFINITY
-1 8.68958 root default
-2 2.17239 host node-4
3 1.08620 osd.3 up 1.00000 1.00000
4 1.08620 osd.4 down 0 1.00000
-3 2.17239 host node-2
2 1.08620 osd.2 up 1.00000 1.00000
6 1.08620 osd.6 up 1.00000 1.00000
-4 2.17239 host node-3
1 1.08620 osd.1 up 1.00000 1.00000
7 1.08620 osd.7 down 0 1.00000
-5 2.17239 host node-1
0 1.08620 osd.0 up 1.00000 1.00000
5 1.08620 osd.5 up 1.00000 1.00000
You have mail in /var/spool/mail/root
[root@node-1 ~]#
检查步骤
登陆对应 osd 节点, 查看对应 osd 服务
[root@node-3 ceph]# systemctl status ceph-osd@7.service
● ceph-osd@7.service - Ceph object storage daemon
Loaded: loaded (/usr/lib/systemd/system/ceph-osd@.service; enabled-runtime; vendor preset: disabled)
Active: failed (Result: start-limit) since Fri 2021-10-29 13:26:40 CST; 4min 11s ago
Process: 184542 ExecStart=/usr/bin/ceph-osd -f --cluster ${CLUSTER} --id %i --setuser ceph --setgroup ceph (code=killed, signal=ABRT)
Process: 184488 ExecStartPre=/usr/lib/ceph/ceph-osd-prestart.sh --cluster ${CLUSTER} --id %i (code=exited, status=0/SUCCESS)
Main PID: 184542 (code=killed, signal=ABRT)
Oct 29 13:26:20 node-3 systemd[1]: Unit ceph-osd@7.service entered failed state.
Oct 29 13:26:20 node-3 systemd[1]: ceph-osd@7.service failed.
Oct 29 13:26:40 node-3 systemd[1]: ceph-osd@7.service holdoff time over, scheduling restart.
Oct 29 13:26:40 node-3 systemd[1]: start request repeated too quickly for ceph-osd@7.service
Oct 29 13:26:40 node-3 systemd[1]: Failed to start Ceph object storage daemon.
Oct 29 13:26:40 node-3 systemd[1]: Unit ceph-osd@7.service entered failed state.
Oct 29 13:26:40 node-3 systemd[1]: ceph-osd@7.service failed.
[root@node-3 ceph]#
重启对应 osd 服务, 无法正常拉起
[root@node-3 ceph]# systemctl restart ceph-osd@7.service
Job for ceph-osd@7.service failed because start of the service was attempted too often. See "systemctl status ceph-osd@7.service" and "journalctl -xe" for details.
To force a start use "systemctl reset-failed ceph-osd@7.service" followed by "systemctl start ceph-osd@7.service" again.
[root@node-3 ceph]#
# 根据提示执行
[root@node-3 ceph]# systemctl reset-failed ceph-osd@7.service
# 执行后重启没有报错，但是状态还是不对，下面需要查日志
[root@node-3 ceph]# systemctl restart ceph-osd@7.service
查看日志, /var/log/ceph-osd@7.log
--- end dump of recent events ---
2021-10-29 13:50:07.117742 7ff76e664800 0 set uid:gid to 167:167 (ceph:ceph)
2021-10-29 13:50:07.117757 7ff76e664800 0 ceph version 10.2.7 (50e863e0f4bc8f4b9e31156de690d765af245185), process ceph-osd, pid 203828
2021-10-29 13:50:07.119187 7ff76e664800 0 pidfile_write: ignore empty --pid-file
2021-10-29 13:50:07.146080 7ff76e664800 0 filestore(/var/lib/ceph/osd/ceph-7) backend xfs (magic 0x58465342)
2021-10-29 13:50:07.146462 7ff76e664800 0 genericfilestorebackend(/var/lib/ceph/osd/ceph-7) detect_features: FIEMAP ioctl is disabled via 'filestore fiemap' config option
2021-10-29 13:50:07.146467 7ff76e664800 0 genericfilestorebackend(/var/lib/ceph/osd/ceph-7) detect_features: SEEK_DATA/SEEK_HOLE is disabled via 'filestore seek data hole' config option
2021-10-29 13:50:07.146483 7ff76e664800 0 genericfilestorebackend(/var/lib/ceph/osd/ceph-7) detect_features: splice is supported
2021-10-29 13:50:07.158475 7ff76e664800 0 genericfilestorebackend(/var/lib/ceph/osd/ceph-7) detect_features: syncfs(2) syscall fully supported (by glibc and kernel)
2021-10-29 13:50:07.158520 7ff76e664800 0 xfsfilestorebackend(/var/lib/ceph/osd/ceph-7) detect_feature: extsize is disabled by conf
2021-10-29 13:50:07.159232 7ff76e664800 1 leveldb: Recovering log #767926
2021-10-29 13:50:07.175968 7ff76e664800 1 leveldb: Delete type=0 #767926
2021-10-29 13:50:07.176019 7ff76e664800 1 leveldb: Delete type=3 #767925
2021-10-29 13:50:07.176880 7ff76e664800 0 filestore(/var/lib/ceph/osd/ceph-7) mount: enabling WRITEAHEAD journal mode: checkpoint is not enabled
2021-10-29 13:50:07.196997 7ff76e664800 -1 journal Unable to read past sequence 1152396918 but header indicates the journal has committed up through 1152397338, journal is corrupt
2021-10-29 13:50:07.270433 7ff76e664800 0 <cls> cls/hello/cls_hello.cc:305: loading cls_hello
2021-10-29 13:50:07.270609 7ff76e664800 0 <cls> cls/cephfs/cls_cephfs.cc:202: loading cephfs_size_scan
2021-10-29 13:50:07.285389 7ff76e664800 0 osd.7 3200 crush map has features 2200130813952, adjusting msgr requires for clients
2021-10-29 13:50:07.285399 7ff76e664800 0 osd.7 3200 crush map has features 2200130813952 was 8705, adjusting msgr requires for mons
2021-10-29 13:50:07.285403 7ff76e664800 0 osd.7 3200 crush map has features 2200130813952, adjusting msgr requires for osds
2021-10-29 13:50:15.583331 7fde2734f800 0 set uid:gid to 167:167 (ceph:ceph)
2021-10-29 13:50:15.583345 7fde2734f800 0 ceph version 10.2.7 (50e863e0f4bc8f4b9e31156de690d765af245185), process ceph-osd, pid 204015
2021-10-29 13:50:15.584786 7fde2734f800 0 pidfile_write: ignore empty --pid-file
2021-10-29 13:50:15.616141 7fde2734f800 0 filestore(/var/lib/ceph/osd/ceph-7) backend xfs (magic 0x58465342)
2021-10-29 13:50:15.616528 7fde2734f800 0 genericfilestorebackend(/var/lib/ceph/osd/ceph-7) detect_features: FIEMAP ioctl is disabled via 'filestore fiemap' config option
2021-10-29 13:50:15.616533 7fde2734f800 0 genericfilestorebackend(/var/lib/ceph/osd/ceph-7) detect_features: SEEK_DATA/SEEK_HOLE is disabled via 'filestore seek data hole' config option
2021-10-29 13:50:15.616550 7fde2734f800 0 genericfilestorebackend(/var/lib/ceph/osd/ceph-7) detect_features: splice is supported
2021-10-29 13:50:15.623022 7fde2734f800 0 genericfilestorebackend(/var/lib/ceph/osd/ceph-7) detect_features: syncfs(2) syscall fully supported (by glibc and kernel)
2021-10-29 13:50:15.623068 7fde2734f800 0 xfsfilestorebackend(/var/lib/ceph/osd/ceph-7) detect_feature: extsize is disabled by conf
2021-10-29 13:50:15.623791 7fde2734f800 1 leveldb: Recovering log #767928
2021-10-29 13:50:15.623840 7fde2734f800 1 leveldb: Level-0 table #767930: started
2021-10-29 13:50:15.628917 7fde2734f800 1 leveldb: Level-0 table #767930: 139 bytes OK
2021-10-29 13:50:15.652033 7fde2734f800 1 leveldb: Delete type=0 #767928
关键日志信息： journal Unable to read past sequence 1152396918 but header indicates the journal has committed up through 1152397338, journal is corrupt
一般强断电会出现这个情况！
处理方法
对应 osd 节点, 修改 ceph 配置文件, 然后重启.
vim /etc/ceph/ceph.conf
[osd]
journal_ignore_corruption = true
# 重启成功后需要注释掉