Re: OSD deadlock with cephfs client and OSD on same machine

[Tommi Virtanen <tv@xxxxxxxxxxx>]

[Whoops, resending as plain text to make vger happy.]

On Fri, Jun 1, 2012 at 2:35 AM, Amon Ott <a.ott@xxxxxxxxxxxx> wrote:
> Thanks for the new log lines in master git. The warning without syncfs()
> support could be a bit more clear though - the system is not only slower, it
> hangs needing a reset and reboot. This is much worse, specially if cephfs is

That warning, introduced in
https://github.com/ceph/ceph/commit/07498d66233f388807a458554640cb77424114c0
, is more about running multiple OSDs on a single server, and without
syncfs(2) one OSD syncing causes all to sync. It's not related to your
case of loopback mounting, what has *never* worked well, with the
apparent exception of ceph-fuse

> say that syncfs() makes it significantly more stable than sync(). We will
> make a several day load test soon.

That still won't make it reliable, just less likely to trigger. Good
luck, you'll need it with loopback mounts.
--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html