On Mon, Dec 26, 2016 at 2:08 AM, xxhdx1985126 <[email protected]> wrote:
>
> According to the following comment, It seems that only when using btrfs which
> would make FileStore to do "parallel journaling" will the replay_version be
> meaningful. Is that right?
>
> // look at whether any of the osds during this interval
> survived
> // past the end of the interval (i.e., didn't crash and
> // potentially fail to COMMIT a write that it ACKed).
There are several different versions in situations related to this; I
haven't looked at the source but I think you're correct about what
this one is.
The other big semi-related one is that we increment a version
("retry_version", I think?) every time we send off an operation
because we haven't gotten a response of any kind and the primary OSD
for an op changed or something.
-Greg
>
>
>
> -------- Forwarding messages --------
> From: "xxhdx1985126" <[email protected]>
> Date: 2016-12-26 16:09:37
> To: "[email protected]" <[email protected]>
> Subject: What is replay_version used for?
>
>
>
> Hi, everyone.
>
>
> What is Objecter::Op::replay_version used for?
>
>
> Thanks:-)
> _______________________________________________
> ceph-users mailing list
> [email protected]
> http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
_______________________________________________
ceph-users mailing list
[email protected]
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
