Documentation #11095
closedUnclear description of when data migration occurs in Adding/Removing OSDs
0%
Description
On a smaller Ceph cluster we had a disk fail.
That probably happened in the middle of the night and nobody noticed.
So the disk had the 'out' state.
During the day, the there was no data-migration active at the time.
Now when you read the documentation:
http://ceph.com/docs/master/rados/operations/add-or-rm-osds/
It clearly says in this order:
- ceph osd out
- Observe the Data Migration
- ceph osd crush remove {name}
So you would get the impression if that happens that there would be no data migration when you remove from the crushmap.
This seems to be not the case.
After you remove a disk from the crush map, there will be a data migration.
Which makes sense if you think about what 'remove from the crushmap' means.
As you can imagine, it's not so much fun on a small cluster to have an unplanned data migration.
Could you please change the documentation to fit the what really happends.
Thank you.
Updated by Leen Besselink about 9 years ago
In theory this could also be a bug. Maybe only in the earlier version we are using now and could have been already fixed.
In that case: could you let me know in what version it was fixed.
It would still be helpful to change the documentation to reflect that this would happen with an earlier version.
Updated by Zac Dover over 4 years ago
- Status changed from New to Closed
This bug has been judged too old to fix. This is because either it is either 1) raised against a version of Ceph prior to Luminous, or 2) just really old, and untouched for so long that it is unlikely nowadays to represent a live documentation concern.
If you think that the closing of this bug is an error, raise another bug of a similar kind. If you think that the matter requires urgent attention, please let Zac Dover know at zac.dover@gmail.com.