[sheepdog] filesystem errors when using sheepdog

Hitoshi Mitake mitake.hitoshi at gmail.com
Fri Jul 17 13:46:13 CEST 2015


At Fri, 17 Jul 2015 20:23:26 +0900,
Hitoshi Mitake wrote:
> 
> At Fri, 17 Jul 2015 14:05:58 +0300,
> Vasiliy Tolstov wrote:
> > 
> > 2015-07-17 8:58 GMT+03:00 Hitoshi Mitake <mitake.hitoshi at gmail.com>:
> > > Vasiliy,
> > >
> > > https://github.com/sheepdog/qemu/tree/serialize-v3
> > >
> > > I updated the patch in the above branch. Could you test it?
> > >
> > > BTW, could you describe your way of reproducing the bug? We are also
> > > trying to reproduce but reproducibility is not so good (we are using
> > > fio). It seems that your testing method has a high reproducibility. If
> > > you can share it, it is a great help.
> > 
> > 
> > This patch works fine if i don't enable discards in qemu. I'm
> > successful install 10 vm in the same time and nothing breaks.
> > But then i enable in qemu discards=unmap i have errors. Does it
> > possible to fix this too?
> 
> Thanks a lot for testing! How about this version for discard:
> https://github.com/sheepdog/qemu/tree/serialize-v4

Vasiliy,

I update the serialize-v4 branch, the branch contains two patches:
https://github.com/sheepdog/qemu/commit/cbce77b5c43da9d24684fc54c6140b50b04a213d
This first one provides a new serialization mechanism, and

https://github.com/sheepdog/qemu/commit/3e8ebf96e981b6dd642db7b60139f76b5dedccf0
this one removes the old serialization mechanism which doesn't work well.

If you can test this branch and if it works well, I can send them to
qemu list. Seems that only your cluster is suitable for this bug,
could you test the branch?

Thanks,
Hitoshi

> 
> Thanks,
> Hitoshi
> 
> > 
> > sheep.log:
> > 
> > Jul 17 13:58:49  ERROR [gway 32830] wait_forward_request(440) fail
> > cccdb100000117, No object found
> > Jul 17 13:58:49  ERROR [io 32792] sd_remove_object(556) failed to
> > remove object cccdb100000117, No object found
> > Jul 17 13:58:49  ERROR [io 32792] update_obj_refcnt(600) fail, 2
> > Jul 17 13:58:49  ERROR [gway 32796] wait_forward_request(440) fail
> > cccdb10000011b, No object found
> > Jul 17 13:58:49  ERROR [io 32791] sd_remove_object(556) failed to
> > remove object cccdb10000011b, No object found
> > Jul 17 13:58:49  ERROR [io 32791] local_discard_obj(969) failed to
> > remove cccdb10000011b
> > Jul 17 13:58:49  ERROR [gway 32757] wait_forward_request(440) fail
> > cccdb100000116, No object found
> > Jul 17 13:58:49  ERROR [io 32789] sd_remove_object(556) failed to
> > remove object cccdb100000116, No object found
> > Jul 17 13:58:49  ERROR [io 32789] local_discard_obj(969) failed to
> > remove cccdb100000116
> > Jul 17 13:58:50  ERROR [gway 32828] wait_forward_request(440) fail
> > cccdb100000119, No object found
> > Jul 17 13:58:50  ERROR [io 32790] sd_remove_object(556) failed to
> > remove object cccdb100000119, No object found
> > Jul 17 13:58:50  ERROR [io 32790] local_discard_obj(969) failed to
> > remove cccdb100000119
> > Jul 17 13:58:50  ERROR [gway 32803] wait_forward_request(440) fail
> > cccdb10000011a, No object found
> > Jul 17 13:58:50  ERROR [io 32795] sd_remove_object(556) failed to
> > remove object cccdb10000011a, No object found
> > Jul 17 13:58:50  ERROR [io 32795] local_discard_obj(969) failed to
> > remove cccdb10000011a
> > Jul 17 13:58:50  ERROR [gway 32823] wait_forward_request(440) fail
> > cccdb100000118, No object found
> > Jul 17 13:58:50  ERROR [io 32793] sd_remove_object(556) failed to
> > remove object cccdb100000118, No object found
> > Jul 17 13:58:50  ERROR [io 32793] update_obj_refcnt(600) fail, 2
> > Jul 17 13:58:50  ERROR [gway 32800] wait_forward_request(440) fail
> > cccdb100000114, No object found
> > Jul 17 13:58:50  ERROR [io 32791] sd_remove_object(556) failed to
> > remove object cccdb100000114, No object found
> > Jul 17 13:58:50  ERROR [io 32791] update_obj_refcnt(600) fail, 2
> > 
> > 
> > -- 
> > Vasiliy Tolstov,
> > e-mail: v.tolstov at selfip.ru


More information about the sheepdog mailing list