radosgw: S3 object retention: high usage of default.rgw.log pool

[Konstantin Shalygin <k0ste@xxxxxxxx>]

Hi.

For some bucket for backup application is applied S3 retention policy, at 04:00 2+days backups will be deleted from bucket.

At this time I see very high usage of default.rgw.log pool. Usage log is enabled, ops log is disabled, index pool on NVMe:

- https://ibb.co/dozqPJ

- https://ibb.co/bMU1Wy

But this usage is not actually usage of hardware (e.g. no io for drives), just 4-5% of ceph-rgw cpu:

- https://ibb.co/nLjPjJ

So what is it - normal behavior or bug? Cluster is running on Luminous 12.2.5.

rgw config diff:

            "rgw_enable_apis": "s3, admin",
            "rgw_enable_usage_log": "true",
            "rgw_frontends": "civetweb port=10.10.10.1:80 num_threads=512",
            "rgw_num_rados_handles": "8",
            "rgw_print_continue": "false",
            "rgw_thread_pool_size": "512",
            "rgw_zone": "default",
            "rgw_zonegroup": "default",

k

_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com