Activity
From 10/05/2020 to 11/03/2020
11/03/2020
- 06:15 PM Bug #48104: fs/ceph: ceph.dir.entries does not acquire necessary caps
- This xattr is populated like this:
return ceph_fmt_xattr(val, size, "%lld", ci->i_files + ci->i_subdirs);
B... - 05:49 PM Bug #48104 (In Progress): fs/ceph: ceph.dir.entries does not acquire necessary caps
- The ceph-mgr asynchronously deletes a directory via libcephfs in the volumes plugin here:...
- 09:31 AM Feature #48057: add debug information for client address
- Have finished it in the debugfs, it will be like:...
- 09:06 AM Feature #48057: add debug information for client address
- Patrick Donnelly wrote:
> Xiubo Li wrote:
> > Patrick Donnelly wrote:
> > > Jeff Layton wrote:
> > > > Given that... - 04:06 AM Feature #48057: add debug information for client address
- Xiubo Li wrote:
> Patrick Donnelly wrote:
> > Jeff Layton wrote:
> > > Given that the idea is to fetch this info f... - 04:05 AM Feature #48057: add debug information for client address
- Patrick Donnelly wrote:
> Jeff Layton wrote:
> > Given that the idea is to fetch this info for a particular mount, ...
11/02/2020
- 03:42 PM Feature #48057: add debug information for client address
- Jeff Layton wrote:
> Given that the idea is to fetch this info for a particular mount, it may be best to implement t... - 02:38 PM Feature #48057: add debug information for client address
- Given that the idea is to fetch this info for a particular mount, it may be best to implement this as a new ioctl().
- 05:20 AM Feature #48057 (In Progress): add debug information for client address
10/30/2020
- 09:31 PM Feature #48057 (Resolved): add debug information for client address
- The ceph-fuse client has a "status" asok command that provides: "inst", "addr", "inst_str", and "addr_str". The "addr...
10/29/2020
- 05:04 PM Feature #48043: add a ->setlease implementation to kcephfs
- I have an implementation that (mostly) works, but a truncate can deadlock the whole thing.
The client issues a tru... - 04:46 PM Feature #48043 (New): add a ->setlease implementation to kcephfs
- We've had a lease (delegation) implementation in libcephfs for some time, and I'd like to add the same to kcephfs, ma...
10/28/2020
- 06:46 AM Bug #45563: __list_add_valid kernel NULL pointer in _ceph_remove_cap
- Jeff Layton wrote:
> I don't think there is much we can do with a kernel this old. Can you reproduce it on something...
10/23/2020
- 04:08 PM Bug #47723 (Pending Backport): corrupted ceph_msg_connect message
- https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=28e1581c3b4ea5f98530064a103c6217bedeea7...
10/20/2020
- 11:59 AM Bug #40340 (Can't reproduce): kernel client stuck at opening forever after network outage.
- No response in over a year. Closing.
- 11:41 AM Bug #45562 (Can't reproduce): soft lockup stuck for 22s! in ceph.ko and code stack is 'destroy_in...
- joe h wrote:
> However,I found a regular phenomenon, When physical node memory is very high(beyond 95%), and excutin...
10/12/2020
- 03:14 AM Bug #46904: kclient: cluster [WRN] client.4478 isn't responding to mclientcaps(revoke)
- Jeff Layton wrote:
> I suspect what's happening is that we end up putting the last reference on some caps, queue the... - 01:19 AM Bug #46904: kclient: cluster [WRN] client.4478 isn't responding to mclientcaps(revoke)
- This is the same issue with https://tracker.ceph.com/issues/47565.
10/09/2020
- 12:00 PM Feature #47808 (New): add "full-force" unmap option
- Currently, there is no way to get rid of the rbd mapping if the cluster becomes inaccessible for a long time or forev...
10/08/2020
- 04:58 PM Bug #47723 (Fix Under Review): corrupted ceph_msg_connect message
- [PATCH] libceph: clear con->out_msg on Policy::stateful_server faults
- 11:35 AM Bug #47723: corrupted ceph_msg_connect message
- I think I see what is going on -- looks like a decade old bug!
10/07/2020
- 02:23 PM Bug #47723: corrupted ceph_msg_connect message
- After staring at this for while, I'm pretty sure that the actual ceph_msg_connect message is fine, but garbage bytes ...
- 11:02 AM Bug #47723: corrupted ceph_msg_connect message
- No, this doesn't ring a bell, and it doesn't look like a cephfs bug. Transferring to libceph.
10/06/2020
- 05:03 PM Bug #47723: corrupted ceph_msg_connect message
- Looks like the client's kernel was 4.18.0-193.el8.x86_64
- 04:43 PM Bug #47723: corrupted ceph_msg_connect message
- Assigning to Ilya for now, though the problem may still be in cephfs code. Any thoughts on this?
- 03:46 PM Bug #47723: corrupted ceph_msg_connect message
- Jeff Layton wrote:
> I see this in the MDS log:
>
> [...]
>
> The version sent is hardcoded to a macro, so thi... - 12:00 PM Bug #47723: corrupted ceph_msg_connect message
- I see this in the MDS log:...
- 11:49 AM Bug #47723: corrupted ceph_msg_connect message
- This is not something I've seen before. What exactly happened in this test?
Looking at /ceph/teuthology-archive/pd...
10/05/2020
- 01:35 PM Bug #47723 (In Progress): corrupted ceph_msg_connect message
- I believe Jeff is looking into this...
Also available in: Atom