I forgot to add some information which is critical here:
On 23/04/18 4:50 PM, Syed Armani wrote:
> Hello folks,
>
> We are running radosgw(Luminous) with Swift API enabled. We observed
> that after updating an object the "hash" and "content_type"
> fields were concatenated with "\u000".
>
> Steps to reproduce the issue.
>
> [1] Create a container(Swift nomenclature)
> [2] Upload a file with "swift --debug upload <name-of-container> <filename>"
> [3] Make a change to the file.
> [4] Update with "swift --debug post <container-name> <filename>"
>
> After the "post" request, the RESP BODY is like this:
We do not see the "\u000" on step 4, instead we see that when we list the
container again after step 4 like this:
[5] swift --debug list <container>
then we see the "\u000" in the output like this:
>
> {
> "bytes": 251,
> "content_type": "text/plain\u0000",
> "hash": "8df6558bcaed354cbf24aa3b4a790fad\u0000",
> "last_modified": "2018-04-23T06:21:52.574Z",
> "name": "test.txt"
> }
Cheers,
Syed
>
> As you can see there is a "\u0000" character in both "content_type" and "hash" fields.
>
>
> Also, when we used the "radosgw-admin object stat" we observed this:
>
> # diff beforeUpdate.txt afterUpdate.txt
> 39c39
> < "tag": "7626b755-713f-4be3-9062-9328ad3fa425.6729.9",
> ---
>> "tag": "_GmhB3jfJWOV7yxkzk0GdrL_8Z71pVN7",
> 79,80c79
> < "user.rgw.tail_tag": "7626b755-713f-4be3-9062-9328ad3fa425.6729.9",
> < "user.rgw.x-amz-meta-mtime": "1524468290.332091"
> ---
>> "user.rgw.tail_tag": "7626b755-713f-4be3-9062-9328ad3fa425.6729.9"
>
> Please note that the "tag" field changed after update and "user.rgw.x-amz-meta-mtime" disappeared.
>
> beforeUpdate.txt --> http://paste.openstack.org/show/719723/
> afterUpdate.txt --> http://paste.openstack.org/show/719724/
>
> We are using a combination of native swift clusters and radsogw based swift clusters.
> Our ambition is to switch to radosgw completely. At the moment issues like above are
> breaking clients which are not able to parse information because of the presence of
> extra "\u0000" character. We donot see any such issue with the native Swift clusters.
>
> Thanks!
>
> Cheers,
> Syed Armani
>
>
_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
