Hi, during upgrade from Octopus (15.2.13) to Pacific (16.2.4) I was not able to complete the upgrade. I've upgraded using the `ceph orch upgrade start` command. After one mon was upgrade, the second upgraded one is not more able to form a quorum, looping following messages in its journal: Aug 12 14:32:04 ceph-node2 bash[4006943]: debug 2021-08-12T12:32:04.680+0000 7f2072ecd700 0 log_channel(cluster) log [INF] : mon.ceph-node2 calling monitor election Aug 12 14:32:04 ceph-node2 bash[4006943]: debug 2021-08-12T12:32:04.680+0000 7f2072ecd700 1 paxos.2).electionLogic(356485) init, last seen epoch 356485, mid-election, bumping Aug 12 14:32:04 ceph-node2 bash[4006943]: debug 2021-08-12T12:32:04.704+0000 7f2072ecd700 1 mon.ceph-node2@2(electing) e12 collect_metadata sda: no unique device id for sda: fallback method has model 'Virtual disk ' but no serial' The cluster has currently three mons, one is upgraded and works, one is not upgraded but also works and the third one loops (and was upgraded). As there is no specific error I do not have any idea how to fix this behavior. As it breaks the work of the cluster itself I've masked the "defective" mon to not start. However this state is everything but sane. Are there some hints how to find the issue? Kind Regards, Michael -- RZ-Status Mitteilungen auf https://twitter.com/WorNetAG ; Michael Wodniok M.Sc. WorNet AG Bürgermeister-Graf-Ring 28 82538 Geretsried Telefon: +49 8171 24691 16 Simply42 und SecuMail sind Marken der WorNet AG. http://www.wor.net/ Handelsregister Amtsgericht München (HRB 129882) Vorstand: Christian Eich Aufsichtsratsvorsitzender: Dirk Steinkopf
_______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
