Hello. I was have one-way multisite S3 cluster and we've seen issues
with rgw-sync due to sharding problems and I've stopped the multisite
sync. This is not the topic just a knowledge about my story.
I have some leftover 0 byte objects in destination and I'm trying to
overwrite them with Rclone "path to path". But somehow I can not
overwrite these objects. If I delete with rclone or rados rm and do
rclone copy again, I got the result below. Rclone gives error but the
object is created again "0 byte" with pending attrs.  Why is this
happening?
I think somehow I need to clean these objects and copy from source
again but how?

What is "user.rgw.olh.pending" ?


[root@SRV1]# radosgw-admin --id radosgw.prod1 object stat
--bucket=mybucket
--object=images/2019/05/29/ad4ba79c-bb66-4ff6-847a-09a1e0cff49f
{
    "name": "images/2019/05/29/ad4ba79c-bb66-4ff6-847a-09a1e0cff49f",
    "size": 0,
    "tag": "713li30rvcrjfwhctx894mj7vf1wa1a8",
    "attrs": {
        "user.rgw.manifest": "",
        "user.rgw.olh.idtag": "v1m9jy4cjck38ptel09qebsbb10pe2af",
        "user.rgw.olh.info": "\u0001\u0001�",
        "user.rgw.olh.pending.00000000606b04728gs23ecq11b3i3l1":
"\u0001\u0001\u0008",
        "user.rgw.olh.pending.00000000606b0472bfhdzxeb9wesd8t7":
"\u0001\u0001\u0008",
        "user.rgw.olh.pending.00000000606b0472fv06t1dob3vmo4da":
"\u0001\u0001\u0008",
        "user.rgw.olh.pending.00000000606b0472lql6c9o88rt211r9":
"\u0001\u0001\u0008",
        "user.rgw.olh.ver": ""
    }
}
[root@SRV1]# rados listxattr -p prod.rgw.buckets.data
c106b26b-xxx-xxxx-xxx-dee3ca5c0968.121384004.3_images/2019/05/29/ad4ba79c-bb66-4ff6-847a-09a1e0cff49f
user.rgw.idtag
user.rgw.olh.idtag
user.rgw.olh.info
user.rgw.olh.ver
[root@SRV1]# rados -p prod.rgw.buckets.data stat
c106b26b-xxx-xxxx-xxx-dee3ca5c0968.121384004.3_images/2019/05/29/ad4ba79c-bb66-4ff6-847a-09a1e0cff49f
prod.rgw.buckets.data/c106b26b-xxx-xxxx-xxx-dee3ca5c0968.121384004.3_images/2019/05/29/ad4ba79c-bb66-4ff6-847a-09a1e0cff49f
mtime 2021-04-05 17:10:55.000000, size 0
_______________________________________________
ceph-users mailing list -- ceph-users@ceph.io
To unsubscribe send an email to ceph-users-le...@ceph.io

Reply via email to