Hi I suggest to upgrade to last Nautilus release! Also, the last Nautilus release doesn't have fix for trimming osdmaps after PG merge [1] (and seems PR's for Nau never be merged). But we push the trimming via restart mon leader 💁♂️
[1] https://github.com/ceph/ceph/pull/43204 k Sent from my iPhone > On 6 Apr 2022, at 21:01, J-P Methot <jp.met...@planethoster.info> wrote: > Hi, > > > On a cluster running Nautilus 14.2.11, the store.db data space usage keeps > increasing. It went from 5GB to 20GB in a year. > > We even had the following warning and adjust ‘mon_data_size_warn’ to 20Gi => > WARNING: MON_DISK_BIG( mon monitor1 is using a lot of disk space ) > > > But the disk space increase is constant about 1.5G per month. > > > We did a 'ceph-monstore-tool /var/lib/ceph/mon/ceph-monitor1/ dump-keys | awk > '{print $1}'| uniq -c’ : > 285 auth > 2 config > 10 health > 1435 logm > 3 mdsmap > 153 mgr > 1 mgr_command_descs > 3 mgr_metadata > 51 mgrstat > 13 mon_config_key > 1 mon_sync > 7 monitor > 1 monitor_store > 5 monmap > 234 osd_metadata > 1 osd_pg_creating > 1152444 osd_snap > 965071 osdmap > 622 paxos > > It appears that the osd_snap is eating up all the space. We have about 1100 > snapshots total (they rotate every 72h). > > I took a look at https://tracker.ceph.com/issues/42012 and it might be > related. However, from the bug report, that particular issue doesn't seem > fixed in Nautilus, but my 14.2.16 cluster that has similar usage doesn't have > this issue. > > > Did anyone face the same issue and do you have a workaround/solution to avoid > mon’s db size increasing constantly ? Could a simple minor version upgrade > fix it or would I need to upgrade to Octopus? > > -- > Jean-Philippe Méthot > Senior Openstack system administrator > Administrateur système Openstack sénior > PlanetHoster inc. > _______________________________________________ > ceph-users mailing list -- ceph-users@ceph.io > To unsubscribe send an email to ceph-users-le...@ceph.io _______________________________________________ ceph-users mailing list -- ceph-users@ceph.io To unsubscribe send an email to ceph-users-le...@ceph.io
