Hi,I've just configured a second zones for 2 of our ceph s3 deployments and I've noticed that after initial sync secondary zone data pool are much bigger than ones on master zones.
My setup consists of main zone, archive zone and sync_policy which configure directional sync from main zone to archive.
here is an example from the zone I've configured today. Master zone looks like thisX.rgw.buckets.data 42 1024 361 GiB 254.26k 542 GiB 0.01 3.1 PiB
secondary archive zone looks like this.X-archive.rgw.buckets.data 32 32 755 GiB 716.68k 1007 GiB 0.03 2.4 PiB
This archive zone was created few hours ago. Users didn't overwrite so much data to double archive zone size. (object count is almost trippled)
I've checked gc list --include-all on archive zones and it's empty. I'm not sure why zone is this big.
Few days ago I've also configured archive zone for different deployment. I've configured archive-zone lifecycle policy to 1 day and tried to cleanup all the buckets on archive zone. It didn't help.
My other archive zone is 150% of the size of it's master zone.I've tried to force sync with `radosgw-admin data sync init` Sync worked but didn't help with excess data on the pool.
I suspect it's an error during initial multisite synchronization. I've restarted RGW daemon on archive zone during initial synchronization in both cases.
What else could have caused this? Are RGW daemons in multisite setups sensitive to restarts?Could similar issues happen during normal rgw restart during multisite operations?
Best regards Adam Prycki
Attachment:
smime.p7s
Description: Kryptograficzna sygnatura S/MIME
_______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
