Hi. Try order o_fs_drbd0_after_ms_drbd0 Mandatory: ms_drbd0:promote fs_drbd0:start
On May 11, 2021 6:35:58 PM Andrei Borzenkov <arvidj...@gmail.com> wrote:
While testing drbd cluster I found errors (drbd device busy) when stopping drbd master with mounted filesystem. I do have order o_fs_drbd0_after_ms_drbd0 Mandatory: ms_drbd0:promote fs_drbd0 and I assumed pacemaker automatically does reverse as "first stop then demote". It does not - umount and demote are initiated concurrently. Adding explicit order o_stop_fs_drbd0_before_demote_ms_drbd0 Mandatory: fs_drbd0:stop ms_drbd0:demote fixed it, but should not this be automatic? The versions is pacemaker-2.0.5+20210310.83e765df6-1.1.x86_64. _______________________________________________ Manage your subscription: https://lists.clusterlabs.org/mailman/listinfo/users ClusterLabs home: https://www.clusterlabs.org/
_______________________________________________ Manage your subscription: https://lists.clusterlabs.org/mailman/listinfo/users ClusterLabs home: https://www.clusterlabs.org/
