Project

General

Profile

Activity

From 01/28/2014 to 02/26/2014

02/25/2014

02:39 AM Bug #7282: Unresponsive rbd-backed Qemu domain causes libvirtd to stall on all connections
The patch has been accepted in libvirt upstream: http://libvirt.org/git/?p=libvirt.git;a=commitdiff;h=60f70542f97805a... Wido den Hollander

02/21/2014

02:11 PM Bug #7385 (In Progress): Objectcacher setting max object counts too low
Sage Weil
10:23 AM Feature #7507 (New): krbd: Make device symlinks cluster aware
Currently when a device is mapped a udev script creates a symlink at /dev/rbd/<pool>/<imagename>
Would be nice if ...
Tyler Brekke

02/19/2014

04:24 AM Bug #6480: librbd crashed qemu-system-x86_64
Just a note, bug itself is very unlikely to reproduce in a short times (one crash per 1k VM instances per month or so). Andrey Korolyov

02/18/2014

05:09 PM Bug #6480 (Need More Info): librbd crashed qemu-system-x86_64
...and now there's a new similar instance, this time on rhel:... Josh Durgin
12:54 PM Bug #6480 (Can't reproduce): librbd crashed qemu-system-x86_64
Ian Colle
02:30 PM Bug #7466 (Resolved): rbd: timeout error not exposed in python bindings
commit:609f4c56718d8279895b02b8163bbe1976c02bfb Josh Durgin
02:21 PM Bug #7466 (Resolved): rbd: timeout error not exposed in python bindings
The recently added timeouts for librados should result in a specific exception type in the python bindings. Josh Durgin
12:55 PM Bug #5488 (Need More Info): librbd: deadlock in image refresh
Ian Colle
12:53 PM Bug #5876 (New): Assertion failure in rbd_img_obj_callback() : rbd_assert(which >= img_request->n...
Ian Colle
12:52 PM Bug #7385: Objectcacher setting max object counts too low
Ian Colle
12:45 PM Bug #7465 (Can't reproduce): krbd: size of disk read or set incorrectly
This was observed on a 3.11 ubuntu kernel:... Josh Durgin

02/17/2014

01:33 PM Bug #7418 (Duplicate): f7cf25f2 is causing KVM processes to use excessive memory
reopened #7385 Sage Weil
01:33 PM Bug #7385 (Pending Backport): Objectcacher setting max object counts too low
This now leaks memory.. presumably stray Object's the cache that never get cleaned up.
Let's just increase the obj...
Sage Weil
01:32 PM Bug #7385: Objectcacher setting max object counts too low
Sage Weil
01:17 PM Feature #7455 (Resolved): krbd,kcephfs: support primary-affinity
when this is complete, be sure to fix ceph-qa-suite.git Sage Weil

02/13/2014

07:41 PM Bug #7418 (Duplicate): f7cf25f2 is causing KVM processes to use excessive memory
Recent change is causing 2GB VM KVM processes to consume up to 5GB of memory during heavy write workload. Sage think... Mark Nelson

02/12/2014

09:54 AM Bug #7385 (Pending Backport): Objectcacher setting max object counts too low
Sage Weil

02/11/2014

11:44 PM Bug #7282: Unresponsive rbd-backed Qemu domain causes libvirtd to stall on all connections
Josh Durgin wrote:
> Hi Wido,
>
> Attached is a patch to use the timeouts with a libvirt storage pool. I noticed ...
Wido den Hollander
06:27 PM Bug #7282: Unresponsive rbd-backed Qemu domain causes libvirtd to stall on all connections
Hi Wido,
Attached is a patch to use the timeouts with a libvirt storage pool. I noticed that libvirt will actually...
Josh Durgin
01:14 PM Bug #7385 (Fix Under Review): Objectcacher setting max object counts too low
https://github.com/ceph/ceph/pull/1215 Josh Durgin
06:07 AM Bug #7385: Objectcacher setting max object counts too low
Does it make sense to be exposing this to librbd/client.cc at all vs just directly setting it via a config option? (o... Mark Nelson

02/10/2014

03:29 PM Bug #7385: Objectcacher setting max object counts too low
This is set by librbd or ceph-fuse's Client.cc after creating the objectcacher. There's already a config option for C... Josh Durgin
03:19 PM Bug #7385 (Resolved): Objectcacher setting max object counts too low
It appears that the objectcacher is setting max object counts based on the max dirty data size and object size. With... Mark Nelson

02/05/2014

03:20 PM Bug #7327: Assertion failure in rbd_object_request_destroy()
No pattern has emerged, except that it happened twice in rapid succession after a long period of not happening (weeks... Nathaniel Filardo
08:28 AM Bug #7327: Assertion failure in rbd_object_request_destroy()
Hi Nathaniel,
What kind of workload are you running? Have you noticed any pattern
wrt when it occurs, what might...
Ilya Dryomov

02/04/2014

05:22 AM Bug #7282: Unresponsive rbd-backed Qemu domain causes libvirtd to stall on all connections
Florian Haas wrote:
> Wido den Hollander wrote:
> > Florian Haas wrote:
> > > Just for clarification: when you say...
Wido den Hollander
05:02 AM Bug #7282: Unresponsive rbd-backed Qemu domain causes libvirtd to stall on all connections
Wido den Hollander wrote:
> Florian Haas wrote:
> > Just for clarification: when you say "it" blocked forever, are ...
Florian Haas
02:34 AM Bug #7282: Unresponsive rbd-backed Qemu domain causes libvirtd to stall on all connections
Florian Haas wrote:
> Wido den Hollander wrote:
> > > * Did you only test with an RBD storage *pool*, or do you als...
Wido den Hollander

02/03/2014

05:30 PM Bug #7327 (Resolved): Assertion failure in rbd_object_request_destroy()
Running Debian kernel 3.12.6 (debian package linux-image-3.12-1-amd64_3.12.6-2_amd64) inside a Xen DomU, we occasiona... Nathaniel Filardo
12:59 PM Bug #7282: Unresponsive rbd-backed Qemu domain causes libvirtd to stall on all connections
Wido den Hollander wrote:
> > * Did you only test with an RBD storage *pool*, or do you also have *domains* running ...
Florian Haas
12:46 PM Bug #7282: Unresponsive rbd-backed Qemu domain causes libvirtd to stall on all connections
Florian Haas wrote:
> Thanks Wido. As explained in the original description we went off 1.1.1 here due to a SIGABRT ...
Wido den Hollander
10:55 AM Bug #7282: Unresponsive rbd-backed Qemu domain causes libvirtd to stall on all connections
Thanks Wido. As explained in the original description we went off 1.1.1 here due to a SIGABRT bug.
Two questions:
...
Florian Haas
06:00 AM Bug #7282: Unresponsive rbd-backed Qemu domain causes libvirtd to stall on all connections
So I've played around with this today and I can't fully reproduce it.
I have a machine running with libvirt 1.1.1-...
Wido den Hollander
01:48 AM Bug #7282: Unresponsive rbd-backed Qemu domain causes libvirtd to stall on all connections
Florian Haas wrote:
> Right, but why should the qemu process stall if all the MONs are available, and only a few RBD...
Wido den Hollander
01:42 AM Bug #7282: Unresponsive rbd-backed Qemu domain causes libvirtd to stall on all connections
Right, but why should the qemu process stall if all the MONs are available, and only a few RBDs are shot? (if they ar... Florian Haas
12:56 AM Bug #7282: Unresponsive rbd-backed Qemu domain causes libvirtd to stall on all connections
I'm going to try to figure this out, I'm setting up a test env right now.
I think it's libvirt which can't handle ...
Wido den Hollander
12:32 AM Bug #7282: Unresponsive rbd-backed Qemu domain causes libvirtd to stall on all connections
With Wido having confirmed that the issue exists, I'm bumping the severity to critical.
Wido, if you could offer y...
Florian Haas
07:25 AM Bug #7310 (Can't reproduce): rbd snap unprotect test failing
... Alfredo Deza

02/02/2014

07:53 AM Bug #7282: Unresponsive rbd-backed Qemu domain causes libvirtd to stall on all connections
Hey Wido,
thanks for responding to this, and doubly for doing so on a weekend. Much appreciated.
> I'm aware of...
Florian Haas

02/01/2014

11:40 AM Bug #7282: Unresponsive rbd-backed Qemu domain causes libvirtd to stall on all connections
I'm aware of this. It's not only with RBD though, the same happens with libvirt when you use NFS and a export is dead... Wido den Hollander

01/31/2014

09:47 AM Bug #7282 (Resolved): Unresponsive rbd-backed Qemu domain causes libvirtd to stall on all connect...
This isn't fully confirmed yet, because we haven't found a reliable way to reproduce. In short, it seems that if you ... Florian Haas
12:47 AM Bug #7076 (Resolved): rbd: check for watchers before removing an image
Ilya Dryomov

01/30/2014

10:16 AM Feature #7272 (Duplicate): rbd: import performance
Currently the rbd import appears to be single threaded which means the import process is being written to a single di... Tyler Brekke

01/29/2014

06:15 AM Bug #7076 (Fix Under Review): rbd: check for watchers before removing an image
wip-rbd-rm-watchers; https://github.com/ceph/ceph/pull/1159 Ilya Dryomov

01/28/2014

11:47 PM Bug #5876: Assertion failure in rbd_img_obj_callback() : rbd_assert(which >= img_request->next_co...
Hi,
I had this error on a 3.10.27 kernel. The fix is included in this kernel, right ?...
Olivier Bonvalet
 

Also available in: Atom