Hello all, Recently I discovered a problem where after taking an OSD out of the cluster and purging it on the OSD node I found that a systemd unit was left over on the system until it was rebooted. This poses a problem in particular for things like prometheus who track these units and think that the OSD is down when it has actually been deleted.
This unit is located in `/run/systemd/system/ceph-osd.target.wants`. The normal process I use to remove the OSD is: * ceph osd purge osd.X --yes-i-really-mean-it * umount /var/lib/ceph/osd/osd-x * ceph-volume lvm zap --destroy /dev/sdX I would have thought that the zap would remove this unit file but it remains on the node until it is rebooted. To me this seems like a bug but perhaps a missed a step from the OSD removal? The below versions of software were used: * ceph - 14.2.18 nautilus * ubuntu - bionic * kernel - 5.4.0-80 Kind Regards, Benard _______________________________________________ ceph-users mailing list -- ceph-users@ceph.io To unsubscribe send an email to ceph-users-le...@ceph.io
