Ceph » Linux kernel client

04/19/2017

03:28 AM Bug #19688 (Fix Under Review): Kernel panic on CephFS kernel client when setting file ACL

fixed by "ceph: fix recursively call between ceph_set_acl and __ceph_setattr" https://github.com/ceph/ceph-client/com... Zheng Yan

01:22 AM Bug #19688 (Resolved): Kernel panic on CephFS kernel client when setting file ACL

From: Jerry Lee <leisurelysw24@gmail.com>... Zheng Yan

04/14/2017

12:25 PM Feature #17204: Implement new-style ENOSPC handling in kclient

I've changed the code in testing to set the barrier first, and added a patch to return writeback errors on close, whi... Jeff Layton

04/13/2017

02:28 PM Feature #17204: Implement new-style ENOSPC handling in kclient

Ok, I see the problem. I added a down_read for the osdc->epoch_barrier fetch to send_cap_msg. That's dumb -- I'll jus... Jeff Layton

02:08 PM Feature #17204: Implement new-style ENOSPC handling in kclient

Ok, hit a deadlock in testing today. Yay for lockdep. I'll have to look at how to fix this:... Jeff Layton

10:22 AM Bug #18690: kclient: FAILED assert(0 == "old msgs despite reconnect_seq feature")

OK, here is the issue reproduced on the kcephfs suite, with the qa/ modification in jcsp/wip-18690-qa
http://pulpi... John Spray

04/12/2017

05:23 PM Bug #18690: kclient: FAILED assert(0 == "old msgs despite reconnect_seq feature")

Thanks for the input Haomai.
Kernel git log has this commit claiming to implement reconnect_seq:... John Spray

04/10/2017

07:59 AM Bug #19309 (Resolved): IO Hang on raw rbd device using libceph kernel module

In 4.4.60, 4.9.21, 4.10.9. Ilya Dryomov

07:57 AM Bug #19309: IO Hang on raw rbd device using libceph kernel module

Discussion with Michal: http://www.spinics.net/lists/stable/msg165052.html Ilya Dryomov

04/07/2017

01:38 PM Bug #18690: kclient: FAILED assert(0 == "old msgs despite reconnect_seq feature")

I suspect simple msgr also has this problem, maybe we can try this Haomai Wang

04/06/2017

06:39 PM Feature #17204: Implement new-style ENOSPC handling in kclient

Ilya seems to be OK with v7 of the set, and I merged that into the kernel testing branch yesterday. How do we turn on... Jeff Layton

04/04/2017

12:50 PM Bug #19127: NULL pointer dereference in ceph_readdir

Jeff Layton wrote:
> Yes, we also don't hold a reference to CEPH_CAP_FILE_SHARED in this code. I suppose that's the ... Zheng Yan

12:31 PM Bug #19127: NULL pointer dereference in ceph_readdir

Jeff Layton wrote:
> So, my thinking was to call ceph_try_get_caps in ceph_readdir to grab CEPH_CAP_FILE_RD and CEPH... Zheng Yan

10:20 AM Bug #18690: kclient: FAILED assert(0 == "old msgs despite reconnect_seq feature")

I read a little at kernel/net/ceph/messenger.cc it looks not respect on reconnect seq actually. maybe we could disabl... Haomai Wang

04/03/2017

02:28 PM Bug #19127: NULL pointer dereference in ceph_readdir

So, my thinking was to call ceph_try_get_caps in ceph_readdir to grab CEPH_CAP_FILE_RD and CEPH_CAP_FILE_SHARED. ceph... Jeff Layton

01:51 PM Bug #19127: NULL pointer dereference in ceph_readdir

Zheng Yan wrote:
> Jeff Layton wrote:
> > Zheng Yan wrote:
> >
> > > no vmcore
> >
> > Bummer. Do you have th... Jeff Layton

01:45 PM Bug #19127: NULL pointer dereference in ceph_readdir

Jeff Layton wrote:
> Zheng Yan wrote:
>
> > no vmcore
>
> Bummer. Do you have the text in the log from before ... Zheng Yan

01:12 PM Bug #19127: NULL pointer dereference in ceph_readdir

Zheng Yan wrote:
> no vmcore
Bummer. Do you have the text in the log from before the Oops line? It'd be goo... Jeff Layton

04/01/2017

03:24 PM Bug #11555 (Resolved): lock inversion related to memory reclaim

We were allocating (with GFP_KERNEL) and destroying the cipher context on each encrypt/decrypt operation:
https://... Ilya Dryomov

03:04 PM Bug #18543 (Closed): rbd map lun02 -p hdd2 rbd: sysfs write failed rbd: map failed: (5) Input/ou...

Ilya Dryomov

10:33 AM Bug #18130 (In Progress): soft lockups in ceph.ko

No, not quite resolved yet. I have a couple of patches for this in the testing branch, but they're marked DNM for now... Jeff Layton

02:44 AM Bug #18130 (Resolved): soft lockups in ceph.ko

relevant code has been removed. (now vfs helpers are used) Zheng Yan

02:50 AM Bug #15432: kcephfs: umount -f can fail after mds reconnect failure

base on Jeff's ENOSPC work, It should be easy to implement function that abort pending osd requests for 'umount -f' Zheng Yan

02:18 AM Bug #19127: NULL pointer dereference in ceph_readdir

Jeff Layton wrote:
> Zheng Yan wrote:
> >
> > we call ceph_dir_clear_ordered() before splice_dentry(). But only f... Zheng Yan

01:11 AM Bug #19127: NULL pointer dereference in ceph_readdir

Jeff Layton wrote:
>Another question too...why are we using CEPH_CAP_FILE_SHARED in this code. Shouldn't we require ... Zheng Yan

03/31/2017

07:17 PM Bug #19419: XFS filesystem on RBD image was corrupt after remount

Hi Ilya,
There were Samba client tests occurring on the platform that were timing out with oplocks. the RBD was be... Bryan Apperson

02:52 PM Bug #19419 (Need More Info): XFS filesystem on RBD image was corrupt after remount

There is no way to tell at this point. It's a possibility, although if that were the case it would have manifested s... Ilya Dryomov

03:29 PM Feature #17524 (In Progress): krbd: support disabling auto-exclusive lock transition logic

Ilya Dryomov

12:02 PM Bug #19127: NULL pointer dereference in ceph_readdir

Zheng Yan wrote:
>
> we call ceph_dir_clear_ordered() before splice_dentry(). But only for dentry's current parent... Jeff Layton

07:41 AM Bug #19127: NULL pointer dereference in ceph_readdir

Jeff Layton wrote:
> We clear the parent's complete bit whenever we call splice_dentry, AFAICT. The only exception i... Zheng Yan

03/30/2017

07:39 PM Bug #19127: NULL pointer dereference in ceph_readdir

We clear the parent's complete bit whenever we call splice_dentry, AFAICT. The only exception is in the readdir code ... Jeff Layton

06:20 PM Bug #19419: XFS filesystem on RBD image was corrupt after remount

the RBD client system was gracefully rebooted at Mar 17 10:22:36. Bryan Apperson

06:13 PM Bug #19419: XFS filesystem on RBD image was corrupt after remount

Bryan Apperson wrote:
> Hi Ilya,
>
> The log starts before the reboot - the corruption appeared before the reboot... Bryan Apperson

06:13 PM Bug #19419: XFS filesystem on RBD image was corrupt after remount

Hi Ilya,
The log starts before the reboot - the corruption appeared before the reboot in the logs. The server had ... Bryan Apperson

04:19 PM Bug #19419: XFS filesystem on RBD image was corrupt after remount

BTW not only firefly is EOL, but that RHEL 7.0 kernel client is also very old. I'd recommend upgrading to the 7.3 ke... Ilya Dryomov

04:09 PM Bug #19419: XFS filesystem on RBD image was corrupt after remount

Hi Bryan,
> Then on March 16th the image became unresponsive. After attempting a reboot of the server and a remoun... Ilya Dryomov

06:09 PM Feature #17204: Implement new-style ENOSPC handling in kclient

Up to v6 of the set now, just re-posted it today. The main difference from v5 is some changes to address Ilya's comme... Jeff Layton

04:01 PM Bug #19385 (Need More Info): rbd stuck on read/write op to ceph_osd

Let me know if you can reproduce this on 4.9.z. Ilya Dryomov

10:07 AM Bug #19122 (Resolved): pre-jewel "osd rm" incrementals are misinterpreted (kernel client)

In 4.4.58, 4.9.19, 4.10.7. Ilya Dryomov

03/29/2017

05:32 PM Bug #19419 (Need More Info): XFS filesystem on RBD image was corrupt after remount

An RBD image was being used in production to back a Samba and NFS export. The 100TB image was formatted with XFS and ... Bryan Apperson

01:07 PM Bug #18690: kclient: FAILED assert(0 == "old msgs despite reconnect_seq feature")

Still seeing this here:
http://pulpito.ceph.com/jspray-2017-03-29_01:19:13-multimds-wip-jcsp-testing-20170328-test... John Spray

03/28/2017

05:03 PM Bug #19385: rbd stuck on read/write op to ceph_osd

Ilya, thanks.
I try it Sergey Jerusalimov

02:33 PM Bug #19385: rbd stuck on read/write op to ceph_osd

There are known issues with resending code in older kernels and manually restarting OSDs is indeed one of the workaro... Ilya Dryomov

11:18 AM Bug #19385: rbd stuck on read/write op to ceph_osd

Ilya, Hi, it's 3.18.43-40 Sergey Jerusalimov

01:41 PM Bug #19127: NULL pointer dereference in ceph_readdir

I suspect the first check does not properly handle the d_splice_alias() case. Zheng Yan

01:39 PM Bug #19127: NULL pointer dereference in ceph_readdir

The idea is borrowed from ncpfs. When receiving reply from server, the readdir code fills dentry pointers in page cac... Zheng Yan

03/27/2017

01:12 PM Bug #19127: NULL pointer dereference in ceph_readdir

Ouch...the ceph readdir code seems to keep pointers to dentries in the ceph_readdir_cache_control, but doesn't hold r... Jeff Layton

01:10 PM Bug #19385: rbd stuck on read/write op to ceph_osd

Which kernel is this on? Ilya Dryomov

09:24 AM Bug #19309 (Pending Backport): IO Hang on raw rbd device using libceph kernel module

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=633ee407b9d15a75ac9740ba9d3338815e1fcb9... Ilya Dryomov

03/24/2017

10:08 PM Bug #19385 (Closed): rbd stuck on read/write op to ceph_osd

Sometimes, after massive osd restart (server with shelf) rbd image (maped and mounted xfs) stuck on read or write ope... Sergey Jerusalimov

03/22/2017

11:13 AM Bug #19309 (Fix Under Review): IO Hang on raw rbd device using libceph kernel module

[PATCH] libceph: force GFP_NOIO for socket allocations Ilya Dryomov

02:12 AM Bug #19309 (In Progress): IO Hang on raw rbd device using libceph kernel module

Ilya Dryomov