Hello,
I have 3 Luminous (12.2.12) clusters setup with rgw multi-site. One cluster
recently suffered data loss from PGs lost to bad SSDs. We're looking at how
to recover from this scenario. The other two clusters are healthy.
Will RGW sync the missing data? Do buckets and objects need to be restored
using rclone?
>From `radosgw-admin sync error list`, these are the error types:
$ cat sync-error-list | jq -r '.[].entries[].info.message' | sort -h |
uniq -c | sort -h
2 failed to sync bucket instance: (13) Permission denied
3 failed to sync object(13) Permission denied
100 failed to sync bucket instance: (125) Operation canceled
239 failed to sync bucket instance: (5) Input/output error
404 failed to read remote metadata entry: (11) Resource temporarily
unavailable
2148 failed to sync object(5) Input/output error
--
Gregory O’Neill
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
