If the fs had to go through journal recovery and reset,
the cephfs-journal-tool resets the journal trim position
because of which the old unused journal objects just stay
forever in the metadata pool. The patch fixes the issue.
Now, the old stale journal objects are trimmed during the
regular trimming cycle helping to recover space in the
metadata pool.
Fixes: https://tracker.ceph.com/issues/69708
Signed-off-by: Kotresh HR <khiremat@redhat.com>
journaler.set_read_pos(new_start);
journaler.set_write_pos(new_start);
journaler.set_expire_pos(new_start);
- journaler.set_trimmed_pos(new_start);
journaler.set_writeable();
cout << "writing journal head" << std::endl;