fixed! thank you for the reply. It was the backslashes in the secret
that was the issue. I generated a new gateway user with:
radosgw-admin user create --uid=test2 --display-name=test2
--access-key={key} --secret={secret_without_slashes} --name
client.radosgw.gateway
and that worked.
On 04/14/2014 09:57 PM, Craig Lewis wrote:
2014-04-14 12:39:20.556085 7f133f7ee700 10 auth_hdr:
GET
x-amz-date:Mon, 14 Apr 2014 11:39:01 +0000
/
2014-04-14 12:39:20.556125 7f133f7ee700 15 *calculated
digest=TQ5LP8ZeufSqKLumak6Aez4o+Pg=*
2014-04-14 12:39:20.556127 7f133f7ee700 15
*auth_sign=hx94rY3BJn7HQKA6ERaksNMQPRs=*
2014-04-14 12:39:20.556127 7f133f7ee700 15 compare=20
2014-04-14 12:39:20.556130 7f133f7ee700 10 *failed to authorize request*
2014-04-14 12:39:20.556167 7f133f7ee700 2 req 2:0.009095:s3:GET
/:list_buckets:http status=403
2014-04-14 12:39:20.556396 7f133f7ee700 1 ====== req done
req=0x8ca280 http_status=403 ======
Did you create all of the rados pools that are mentioned in ceph.conf
and the region and zone maps?
The hash s3cmd computed is different that the one RGW computed.
Can you verify that the access and secret keys in .s3cfg match the
output of:
radosgw-admin user info --uid=test1 --name client.radosgw.gateway
Does the secret have a backslash (\) character in it? The docs warn
that not everything handles it well. I regenerated my keys, rather
than testing if s3cmd worked correctly.
_______________________________________________
ceph-users mailing list
[email protected]
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
--
Systems and Storage Engineer, Digital Repository of Ireland (DRI)
High Performance & Research Computing, IS Services
Lloyd Building, Trinity College Dublin, Dublin 2, Ireland.
http://www.tchpc.tcd.ie/ | [email protected]
Tel: +353-1-896-4466
_______________________________________________
ceph-users mailing list
[email protected]
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
