[sheepdog-users] Testing snapshot-object-reclaim

[Liu Yuan]

On Thu, Mar 06, 2014 at 01:34:34PM +0900, Hitoshi Mitake wrote:
> At Thu, 6 Mar 2014 11:19:47 +0800,
> Liu Yuan wrote:
> > 
> > On Wed, Mar 05, 2014 at 03:31:21PM +0100, Valerio Pachera wrote:
> > > 2014-03-05 12:52 GMT+01:00 Valerio Pachera <sirio81 at gmail.com>:
> > > > I'll repeat the test and copy / paste the error.
> > > 
> > > 
> > > This time I was connected by ssh and dd got stuck.
> > > Guest vnc is showing the messages attached.
> > > 
> > > In sheep.log I got errors
> > > 
> > > Mar 05 12:26:53  ERROR [gway 6289] wait_forward_request(438) fail
> > > 7c2b270000025b, No object found
> > > Mar 05 12:26:53  ERROR [gway 6289] wait_forward_request(438) fail
> > > 7c2b270000025b, No object found
> > > Mar 05 12:26:53  ERROR [gway 6289] wait_forward_request(438) fail
> > > 7c2b270000025b, No object found
> > > Mar 05 12:26:53  ERROR [io 6243] sd_remove_object(451) failed to
> > > remove object 7c2b270000025b, No object found
> > > Mar 05 12:26:53  ERROR [io 6243] peer_decref_object(1116) error No object found
> > > Mar 05 12:26:59  ERROR [gway 6281] wait_forward_request(438) fail
> > > 7c2b2700000420, No object found
> > > Mar 05 12:26:59  ERROR [gway 6281] wait_forward_request(438) fail
> > > 7c2b2700000420, No object found
> > > Mar 05 12:26:59  ERROR [gway 6281] wait_forward_request(438) fail
> > > 7c2b2700000420, No object found
> > > Mar 05 12:26:59  ERROR [io 6136] sd_remove_object(451) failed to
> > > remove object 7c2b2700000420, No object found
> > > Mar 05 12:26:59  ERROR [io 6136] peer_decref_object(1116) error No object found
> > > Mar 05 12:27:11  ERROR [gway 6281] wait_forward_request(438) fail
> > > 7c2b270000082a, No object found
> > > Mar 05 12:27:11  ERROR [gway 6281] wait_forward_request(438) fail
> > > 7c2b270000082a, No object found
> > > Mar 05 12:27:11  ERROR [gway 6281] wait_forward_request(438) fail
> > > 7c2b270000082a, No object found
> > > Mar 05 12:27:11  ERROR [io 6134] sd_remove_object(451) failed to
> > > remove object 7c2b270000082a, No object found
> > > Mar 05 12:27:11  ERROR [io 6134] peer_decref_object(1116) error No object found
> > > Mar 05 12:43:01  ERROR [gway 6352] wait_forward_request(438) fail
> > > 87c2b2800000008, No object found
> > > Mar 05 12:43:01  ERROR [gway 6352] wait_forward_request(438) fail
> > > 87c2b2800000008, No object found
> > > Mar 05 12:43:01  ERROR [gway 6349] sd_dec_object_refcnt(492) failed to
> > > decrement reference 87c2b2800000008, No object found
> > > Mar 05 12:43:01  ERROR [gway 6349] update_obj_refcnt(603) fail, 2
> > > Mar 05 12:43:37  ERROR [gway 6397] wait_forward_request(438) fail
> > > 87c2b2800000009, No object found
> > > Mar 05 12:43:37  ERROR [gway 6398] wait_forward_request(438) fail
> > > 87c2b2800000121, No object found
> > > Mar 05 12:43:37  ERROR [gway 6397] wait_forward_request(438) fail
> > > 87c2b2800000009, No object found
> > > Mar 05 12:43:37  ERROR [gway 6375] sd_dec_object_refcnt(492) failed to
> > > decrement reference 87c2b2800000009, No object found
> > > Mar 05 12:43:37  ERROR [gway 6375] update_obj_refcnt(603) fail, 2
> > > Mar 05 12:43:37  ERROR [gway 6398] wait_forward_request(438) fail
> > > 87c2b2800000121, No object found
> > > Mar 05 12:43:37  ERROR [gway 6391] sd_dec_object_refcnt(492) failed to
> > > decrement reference 87c2b2800000121, No object found
> > > Mar 05 12:43:37  ERROR [gway 6391] update_obj_refcnt(603) fail, 2
> > > Mar 05 12:43:37  ERROR [gway 6343] wait_forward_request(438) fail
> > > 87c2b2800000220, No object found
> > > Mar 05 12:43:37  ERROR [gway 6343] wait_forward_request(438) fail
> > > 87c2b2800000220, No object found
> > > Mar 05 12:43:37  ERROR [gway 6399] sd_dec_object_refcnt(492) failed to
> > > decrement reference 87c2b2800000220, No object found
> > 
> > 
> > > -- 
> > 
> > Hi Valerio,
> > 
> > Seems that you still have some problem with the laest reclaim branch?
> 
> The latest reclaim branch wouldn't be able to pass the Valerio's
> test. I'll work on it soon.
> 
> But the problem is a the GC algorithm specific thing, because the
> branch can pass all the tests. It doesn't cause degradation of
> existing sheepdog functionalities.

Since for now there are no major conflicting patches around and I think even
in a future window, no conficting patches against your patch set, since Robin
and I will mainly work on http storage, how about we merge your patch set after
you fixing Valerio's problem?

If there are conflicting patchest coming around, then we can consider merge
your patch set to move on easily.

Thanks
Yuan