Bug #4713
closed
- Subject changed from mds: ha to mds: hang related to access from two clients
I discovered this while trying to reproduce the issue
in http://tracker.ceph.com/issues/4706.
I documented it there, but I'm pretty sure it's a
quite different problem from that one.Specifically what I'm doing:
Here's how to reproduce it:
machine 1
---------
hostname > /mnt/ceph/foo
cat /mnt/ceph/foo
(shows hostname)
machine 2
---------
cat /mnt/ceph/foo
(shows machine 1 hostname)
hostname > /mnt/ceph/foo
cat /mnt/ceph/foo
(shows machine 2 hostname,
then hangs)
cat /mnt/ceph/foo
(no output, hangs)
I got this running this for ceph-client (current "testing"):
51122dd libceph: make method call data be a separate data item
and this for ceph (the current "bobtail" branch) (or close):
daa6ed2 Merge pull request #210 from dalgaaf/wip-da-bobtail-pybind
- Status changed from New to Fix Under Review
- Assignee set to Alex Elder
I have tested the commands listed above on a system with the
patches described here:
http://tracker.ceph.com/issues/4679
I no longer see the hang.
This issue can be marked resolves as soon as those are committed.
Assigning this to myself as a reminder to do this.
- Category set to 47
- Status changed from Fix Under Review to Resolved
- Target version set to v0.62a
The following have been committed to the "testing" branch
of the ceph-client git repository. With them in place
I no longer see this hang.
625d6ec ceph: fix race between writepages and truncate
8bc726f ceph: apply write checks in ceph_aio_write
e7a4c3e ceph: take i_mutex before getting Fw cap
Also available in: Atom
PDF