Hi,

On 08.08.25 15:37, Robert Sander wrote:

As can be seen a third DB volume has been added to sdc
and the OSD uses all of the 18TB of sdf.

When adding a completely new disk a fourth DB volume gets created on sdb
and the new OSD on sdl uses all of the 18TB.

Something must have been changed between reef and squid, not for the better.

No I changed the drivegroup spec to explicitely name the DB devices:

service_type: osd
service_id: hdd2025
service_name: osd.hdd2025
placement:
  label: osd2025
spec:
  block_db_size: 100GB
  data_devices:
    rotational: true
    size: '16TB:'
  db_devices:
    paths:
    - /dev/sdb
    - /dev/sdc

This still works as expected on 18.2.7: adding a new 18TB HDD adds another DB volume on sdb and a new OSD.

On 19.2.3 the orchestrator now does nothing. It does not touch the new HDD (which is listed as available in "ceph orch device ls"). It does not create a DB volume on the HDD (as before) or on the SSD as instructed.

The is a real bad regression.

With this behaviour of the orchestrator I am not able to recommend Ceph 19 to my customers. Appr half of the run Ceph on HDD with the DB volumes on external drives.

Adding or replacing disks becomes a manual task again including creating the DB volumes. This was definitely better in Ceph 18.

Regards
--
Robert Sander
Heinlein Consulting GmbH
Schwedter Str. 8/9b, 10119 Berlin

http://www.heinlein-support.de

Tel: 030 / 405051-43
Fax: 030 / 405051-19

Zwangsangaben lt. §35a GmbHG:
HRB 220009 B / Amtsgericht Berlin-Charlottenburg,
Geschäftsführer: Peer Heinlein -- Sitz: Berlin
_______________________________________________
ceph-users mailing list -- ceph-users@ceph.io
To unsubscribe send an email to ceph-users-le...@ceph.io

Reply via email to