Radosgw: object lifecycle (expiration) not working?

[Robert Stanford <rstanford8896@xxxxxxxxx>]

 Greetings -

 I have created several test buckets in radosgw, to test different expiration durations:

 $ s3cmd mb s3://test2d

 I set a lifecycle for each of these buckets:

 $ s3cmd setlifecycle lifecycle2d.xml s3://test2d --signature-v2

 The files look like this:

<?xml version="1.0" encoding="UTF-8"?>
<LifecycleConfiguration xmlns="http://s3.amazonaws.com/doc/2006-03-01/"><Rule><ID>test2d</ID><Status>Enabled</Status><Expiration><Days>2</Days></Expiration><Prefix></Prefix></Rule></LifecycleConfiguration>

 I tried different ways of applying the lifecycle with s3cmd, and the approach I've used above is the only one that doesn't return a 'feature not implemented' (or similar) error message.

 I've uploaded a file to each bucket, and waited.  Several days have passed.  The files have *not* been removed from any of the buckets, though several of the lifecycle passes have run:

ceph01# radosgw-admin lc list
[
    {
        "bucket": ":stestbucket_1d:bc186feb-9fd2-4035-aa91-9ec0772eefeb.23420.1",
        "status": "COMPLETE"
    },
    {
        "bucket": ":stestbucket_2d:bc186feb-9fd2-4035-aa91-9ec0772eefeb.23420.2",
        "status": "COMPLETE"
    },
    {
        "bucket": ":stestbucket_5d:bc186feb-9fd2-4035-aa91-9ec0772eefeb.23420.3",
        "status": "COMPLETE"
    },
    {
        "bucket": ":stestbucket_7d:bc186feb-9fd2-4035-aa91-9ec0772eefeb.23420.4",
        "status": "COMPLETE"
    },
    {
        "bucket": ":t_testbucket_1d:bc186feb-9fd2-4035-aa91-9ec0772eefeb.40904.1",
        "status": "COMPLETE"
    },
    {
        "bucket": ":t_testbucket_2d:bc186feb-9fd2-4035-aa91-9ec0772eefeb.40904.2",
        "status": "COMPLETE"
    },
    {
        "bucket": ":t_testbucket_5d:bc186feb-9fd2-4035-aa91-9ec0772eefeb.40904.3",
        "status": "COMPLETE"
    },
    {
        "bucket": ":t_testbucket_7d:bc186feb-9fd2-4035-aa91-9ec0772eefeb.40904.4",
        "status": "UNINITIAL"
    },
    {
        "bucket": ":testbucket:bc186feb-9fd2-4035-aa91-9ec0772eefeb.5492.2",
        "status": "UNINITIAL"
    }
]

  The UNINITIAL status entries haven't run yet (not enough time has passed, in the case of 7d).

An s3cmd ls confirms that the files are still in all the buckets.

  Additional relevant information: this is my second test run.  The first time, object expiration seemed to work... except that all objects were deleted at the first run, despite having a different "Days" number in the XML.  I thought this might be due to the "ID" in the XML being the same for each, so for the second test run (which you see above) I changed the "ID" field as well as "Days".

 Anyone know what's going on here?

 Regards and thanks

 RS

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