Project

General

Profile

Activity

From 12/22/2016 to 01/20/2017

01/20/2017

03:24 PM Feature #17204: Implement new-style ENOSPC handling in kclient
v1 of the patch series:
http://marc.info/?l=ceph-devel&m=148492546411549&w=2
Jeff Layton
11:36 AM Bug #18161 (Resolved): kernel client failing to look up mds_namespace gives ENOENT (but it exists)
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=cc8e8342930129aa2c9b629e1653e4681f0896ea i... Ilya Dryomov
10:41 AM Feature #4690: krbd: support arbitrary length responses to class operations
No, just some groundwork. Ilya Dryomov
10:39 AM Bug #18543: rbd map lun02 -p hdd2 rbd: sysfs write failed rbd: map failed: (5) Input/output error
It looks like you have header CRCs disabled ("ms crc header = false" in ceph.conf). This is not supported by the ker... Ilya Dryomov

01/19/2017

11:58 PM Feature #4690: krbd: support arbitrary length responses to class operations
@Ilya: was this already completed? Jason Dillaman

01/18/2017

09:52 PM Feature #17204: Implement new-style ENOSPC handling in kclient
Ok, and now I think I see part of the fundamental issue and that may not have been a problem with the older request h... Jeff Layton
07:49 PM Feature #17204: Implement new-style ENOSPC handling in kclient
Ok, opened up a bug to track the problem I was seeing. dis and sage set me right on IRC though. John's original serie... Jeff Layton

01/17/2017

09:34 PM Feature #17204: Implement new-style ENOSPC handling in kclient
Evidently the OSD code sets the flags to -1 on incremental map updates?... Jeff Layton
08:52 PM Feature #17204 (In Progress): Implement new-style ENOSPC handling in kclient
I've been steadily working on this, and am seeing some weird behavior that I'm still trying sort out. Sometimes, it t... Jeff Layton
08:32 AM Bug #18543 (Closed): rbd map lun02 -p hdd2 rbd: sysfs write failed rbd: map failed: (5) Input/ou...

[root@node03 ceph]# rbd map lun02 -p hdd2
rbd: sysfs write failed
rbd: map failed: (5) Input/output error
...
rook ker

01/16/2017

05:56 PM Feature #17204: Implement new-style ENOSPC handling in kclient
Yeah, I'll get there eventually on the testcases. For now I'm just doing "by hand" testing to work out the problems i... Jeff Layton

01/14/2017

12:25 AM Bug #18161: kernel client failing to look up mds_namespace gives ENOENT (but it exists)
Should undo https://github.com/ceph/ceph-qa-suite/pull/1156/commits/5f1abf9c310c2732cc6bcd0ff2bd2e947dfb414e when thi... Greg Farnum

01/13/2017

07:04 AM Feature #17805 (In Progress): Match fuse_require_active_mds behaviour in kernel client
Ping John and Zheng. Greg Farnum
06:57 AM Feature #17204: Implement new-style ENOSPC handling in kclient
We also need to undo https://github.com/ceph/ceph-qa-suite/pull/1156/commits/c5f6dfc14f47cca251dcac5c53f6369fd36ace1a... Greg Farnum

01/12/2017

07:25 PM Feature #17204: Implement new-style ENOSPC handling in kclient
This was quite a while ago. But generally, the OSDs don't send back errors; they just block. We also want to return E... Greg Farnum
07:14 PM Feature #17204: Implement new-style ENOSPC handling in kclient
Ok, I've looked over the kernel patches and started backporting them, and also on the userland equivalents. There has... Jeff Layton
03:21 PM Bug #17410 (Closed): rbd image stale/stuck (mapped and mounted)
This bug is fixed in kernels 4.7 and above (and also in RHEL 7.3 based kernels, e.g. kernel-3.10.0-514.2.2.el7).
T...
Ilya Dryomov
03:00 PM Bug #2439 (Resolved): ceph-client: auth: handle null verify_authorizer_reply method
All the plumbing for verify_authorizer_reply() is there, but it wasn't being called by the messenger: https://git.ker... Ilya Dryomov
01:30 PM Bug #18148 (Resolved): Fix unsafe d_parent handling in kcephfs
by commits c3f4688a08fd86f1bf8e055724c84b7a40a09733
Zheng Yan

01/10/2017

04:59 PM Bug #18130: soft lockups in ceph.ko
New patch sent this morning that seems to fix this problem. Still waiting for feedback from Al Viro as to why ITER_BV... Jeff Layton
12:40 PM Bug #18474: oops in __unregister_request
On the next attempt:... Jeff Layton
12:15 PM Bug #18474 (Resolved): oops in __unregister_request
I left a xfstests run going overnight last night and when I came back I saw this oops on the console:... Jeff Layton

01/06/2017

04:09 PM Bug #18130: soft lockups in ceph.ko
Bulding and testing a revised patch now.
iov_iter_alignment is not quite right since it takes the iov_len into acc...
Jeff Layton

01/05/2017

05:06 PM Bug #18130: soft lockups in ceph.ko
Ok, I think the problem is actually here and this patch fixes it:... Jeff Layton
03:08 PM Bug #18130: soft lockups in ceph.ko
fio jobfile that easily reproduces this:... Jeff Layton
01:36 PM Bug #18130: soft lockups in ceph.ko
Ok, revamped my debugging a bit and fixed the code to error out when no progress is being made. Now I see this:
<p...
Jeff Layton

01/04/2017

08:30 PM Bug #18130: soft lockups in ceph.ko
Added a bit more info, including the tgid for the task. That allows us to tell what thread context each of these is o... Jeff Layton
06:07 PM Bug #18130: soft lockups in ceph.ko
I think I see the problem. Still looking at the fix. I converted some dout messages to printks (just for simplicity) ... Jeff Layton
05:17 PM Bug #18130: soft lockups in ceph.ko
Ok, this is proving difficult to troubleshoot as the softlockups end up bringing the box to its knees.
I think at ...
Jeff Layton
12:27 AM Bug #18130: soft lockups in ceph.ko
I don't see where ITER_PIPE gets used in the write codepaths. Maybe ITER_BVEC needs special handling? In any case, I'... Jeff Layton
08:33 AM Bug #18161: kernel client failing to look up mds_namespace gives ENOENT (but it exists)
https://github.com/ceph/ceph-client/commit/9c0e972bd7c467f1e70f123372e492404580296c Zheng Yan
 

Also available in: Atom