You can't downgrade from Luminous to Kraken.... well officially at least. I guess it maybe could somehow work but you'd need to re-create all the services. For the mon example: delete a mon, create a new old one, let it sync, etc. Still a bad idea.
Paul -- Paul Emmerich Looking for help with your Ceph cluster? Contact us at https://croit.io croit GmbH Freseniusstr. 31h 81247 München www.croit.io Tel: +49 89 1896585 90 On Wed, Aug 21, 2019 at 1:37 PM nokia ceph <nokiacephus...@gmail.com> wrote: > > Hi Team, > > One of our old customer had Kraken and they are going to upgrade to Luminous > . In the process they also requesting for downgrade procedure. > Kraken used leveldb for ceph-mon data , from luminous it changed to rocksdb , > upgrade works without any issues. > > When we downgrade , the ceph-mon does not start and the mon kv_backend not > moving away from rocksdb . > > After downgrade , when kv_backend is rocksdb following error thrown by > ceph-mon , trying to load data from rocksdb and end up in this error, > > 2019-08-21 11:22:45.200188 7f1a0406f7c0 4 rocksdb: Recovered from manifest > file:/var/lib/ceph/mon/ceph-cn1/store.db/MANIFEST-000716 > succeeded,manifest_file_number is 716, next_file_number is 718, last_sequence > is 311614, log_number is 0,prev_log_number is 0,max_column_family is 0 > > 2019-08-21 11:22:45.200198 7f1a0406f7c0 4 rocksdb: Column family [default] > (ID 0), log number is 715 > > 2019-08-21 11:22:45.200247 7f1a0406f7c0 4 rocksdb: EVENT_LOG_v1 > {"time_micros": 1566386565200240, "job": 1, "event": "recovery_started", > "log_files": [717]} > 2019-08-21 11:22:45.200252 7f1a0406f7c0 4 rocksdb: Recovering log #717 mode 2 > 2019-08-21 11:22:45.200282 7f1a0406f7c0 4 rocksdb: Creating manifest 719 > > 2019-08-21 11:22:45.201222 7f1a0406f7c0 4 rocksdb: EVENT_LOG_v1 > {"time_micros": 1566386565201218, "job": 1, "event": "recovery_finished"} > 2019-08-21 11:22:45.202582 7f1a0406f7c0 4 rocksdb: DB pointer 0x55d4dacf0000 > 2019-08-21 11:22:45.202726 7f1a0406f7c0 -1 ERROR: on disk data includes > unsupported features: compat={},rocompat={},incompat={9=luminous ondisk > layout} > 2019-08-21 11:22:45.202735 7f1a0406f7c0 -1 error checking features: (1) > Operation not permitted > > We changed the kv_backend file inside /var/lib/ceph/mon/ceph-cn1 to leveldb > and ceph-mon failed with following error, > > 2019-08-21 11:24:07.922978 7fc5a25de7c0 -1 WARNING: the following dangerous > and experimental features are enabled: bluestore,rocksdb > 2019-08-21 11:24:07.922983 7fc5a25de7c0 0 set uid:gid to 167:167 (ceph:ceph) > 2019-08-21 11:24:07.923009 7fc5a25de7c0 0 ceph version 11.2.0 > (f223e27eeb35991352ebc1f67423d4ebc252adb7), process ceph-mon, pid 3509050 > 2019-08-21 11:24:07.923050 7fc5a25de7c0 0 pidfile_write: ignore empty > --pid-file > 2019-08-21 11:24:07.944867 7fc5a25de7c0 -1 WARNING: the following dangerous > and experimental features are enabled: bluestore,rocksdb > 2019-08-21 11:24:07.950304 7fc5a25de7c0 0 load: jerasure load: lrc load: isa > 2019-08-21 11:24:07.950563 7fc5a25de7c0 -1 error opening mon data directory > at '/var/lib/ceph/mon/ceph-cn1': (22) Invalid argument > > Is there any possibility to toggle ceph-mon db between leveldb and rocksdb? > Tried to add mon_keyvaluedb = leveldb and filestore_omap_backend = leveldb in > ceph.conf also not worked. > thanks, > Muthu > _______________________________________________ > ceph-users mailing list > ceph-users@lists.ceph.com > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com _______________________________________________ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
