Hi,
Am 05.05.21 um 13:39 schrieb Joachim Kraftmayer:
the crush rule with ID 1 distributes your EC chunks
over the osds
without considering the ceph host. As Robert already suspected.
Yes, the "nxtcloudAF" rule is not fault tolerant enough. Having the OSD
as failure zone will lead to data loss or at least intermediate
unavailability.
The situation is now that all copies (resp. EC chunks) for a PG are
stored on OSDs of the same host. These PGs will be unavailable if the
host is down.
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 93818 B / Amtsgericht Berlin-Charlottenburg,
Geschäftsführer: Peer Heinlein -- Sitz: Berlin