Hello，everyone,
In the OS disk blocked scene, the Mon service is still running ,but cant't work normally,
soon the mon will out of quorum, but some OSD were still markd down after mon_osd_report_timeout*2 seconds,
which will cause the cluster unavailable.
At this time， the OS is very slow，may be unable to operate or login，so it‘s impossible to kill the mon by hand or other service.
This problem exist in L and N version.
Any suggestion？
you can reproduce this problem as follow:
1、 block the OS disk use follow command.
# echo blocked > /sys/block/sdx/device/state
2、a few seconds later， the mon will out of quorum
3、after mon_osd_report_timeout*2 seconds，some osds will be markd down.
（mon_osd_report_timeout default value 900s，you can set a smaller value）
912273695@xxxxxx
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
