Project

General

Profile

Activity

From 11/12/2018 to 12/11/2018

12/11/2018

09:13 PM Bug #37486: tmpfs in /var/lib/ceph/osd/X sometimes created with wrong permissions
master PR https://github.com/ceph/ceph/pull/25477 Alfredo Deza
03:14 PM Bug #37449: zap: improve zapping to remove all partitions and all LVs, encrypted or not
master PR: https://github.com/ceph/ceph/pull/25330 Alfredo Deza
01:20 PM Bug #37595 (In Progress): device_id function causes breakage when trying to format
master PR https://github.com/ceph/ceph/pull/25469 Alfredo Deza
01:18 PM Bug #37595 (Resolved): device_id function causes breakage when trying to format
... Alfredo Deza
08:52 AM Bug #37590 (Resolved): api.vgcreate uses a PE size of 1G
Using such a large extend size leads to potentially unused space of just under 1G per devices.
Why was such a large ...
Jan Fajerski

12/10/2018

06:10 PM Bug #37486: tmpfs in /var/lib/ceph/osd/X sometimes created with wrong permissions
Potential working fix:... Alfredo Deza
04:34 PM Bug #37486 (In Progress): tmpfs in /var/lib/ceph/osd/X sometimes created with wrong permissions
... Alfredo Deza
04:44 PM Bug #37589 (New): argument validators can eat up exceptions
If an error happens inside an argument validator, the exception is eaten up resulting in a cryptic error message
<...
Alfredo Deza
01:03 PM Bug #36492: ceph lvm list reports wrong json from within a container
Ok, no problem.
Thx for your answer Alfredo. For the moment, I have built "my" Mimic ceph/daemon docker container ...
Francois Lafont
11:57 AM Bug #36492: ceph lvm list reports wrong json from within a container
Correct, there hasn't been a Mimic release that includes the fixes here. You will have to wait for one, hopefully soon! Alfredo Deza

12/07/2018

09:43 PM Bug #37486: tmpfs in /var/lib/ceph/osd/X sometimes created with wrong permissions
13.2.2 on Debian Stretch.
The problem was that root owns . and ceph the symlink, in your example root owns everyth...
Paul Emmerich
09:24 PM Bug #37486: tmpfs in /var/lib/ceph/osd/X sometimes created with wrong permissions
What version of Ceph was this? There was a problem with permissions a while ago, see issue: http://tracker.ceph.com/i... Alfredo Deza

12/05/2018

10:35 AM Bug #37506: Change in util/disk.py get_devices() breaks inventory
https://github.com/ceph/ceph/pull/25390 Jan Fajerski
01:08 AM Bug #36492: ceph lvm list reports wrong json from within a container
Hi @all. :)
I'm trying to deploy a Ceph Mimic cluster in docker containers via ceph-ansible and I'm currently stuc...
Francois Lafont

12/04/2018

03:24 PM Bug #37506: Change in util/disk.py get_devices() breaks inventory
inventory breaks because it tries to access the field 'human_readable_size' which the new 'partition devices' don't h... Jan Fajerski
12:12 PM Bug #37502: lvm batch potentially creates multi-pv volume groups
That is correct, we did try to implement this with one VG per backing device and it was incredibly difficult. Doing a... Alfredo Deza

12/03/2018

03:38 PM Bug #37506 (Resolved): Change in util/disk.py get_devices() breaks inventory
This commit https://github.com/ceph/ceph/commit/7f3c3598491e556ec1b93ea0a54f52144202fab6
breaks the inventory func...
Jan Fajerski
02:35 PM Bug #37504 (Can't reproduce): TestValidDevice tests are failing
... Jan Fajerski
10:13 AM Feature #37083 (Resolved): ceph-volume should report device_id as part of the device list
https://github.com/ceph/ceph/pull/25201 Jan Fajerski
09:48 AM Bug #37502: lvm batch potentially creates multi-pv volume groups
I guess the advantage of the current implementation is that a single vg is easier to manage? Jan Fajerski
09:47 AM Bug #37502 (Resolved): lvm batch potentially creates multi-pv volume groups
Both bluestore and filestore MixedStrategy create one volume group if multiple free SSDs are detected. This can creat... Jan Fajerski
09:41 AM Bug #37488 (Resolved): udevadm parser needs a strict mode
Fixed by https://github.com/ceph/ceph/pull/25201 Jan Fajerski

11/30/2018

08:09 PM Bug #37492 (New): ceph-volume lvm is unreliable if the system contains broken disks
Almost all @ceph-volume lvm@ commands invoke some lvm command which access *all* disks. A simple example is just @lvs... Paul Emmerich
06:47 PM Bug #37490 (Resolved): ceph-volume lvm list is O(n^2)
@ceph-volume lvm list@ calls @/sbin/lvs --noheadings --readonly --separator=";" -o lv_tags,lv_path,lv_name,vg_name,lv... Paul Emmerich
03:17 PM Bug #37488 (Resolved): udevadm parser needs a strict mode
Once https://github.com/ceph/ceph/pull/25330 gets merged, there is potential for unsafe parsing getting ignored. A st... Alfredo Deza
02:13 PM Bug #37487 (New): many ceph-volume commands give up after one disk fails with some error
For example, @ceph-volume lvm activate --all@ stops if it encounters an error with any of the disks.
This can make...
Paul Emmerich
02:07 PM Bug #37486: tmpfs in /var/lib/ceph/osd/X sometimes created with wrong permissions
Alfredo Deza
02:01 PM Bug #37486 (Resolved): tmpfs in /var/lib/ceph/osd/X sometimes created with wrong permissions
I've encountered a strange issue on a customer's system today: Some OSDs didn't come up after a reboot, it was somewh... Paul Emmerich
09:24 AM Bug #37442: ceph-volume: Unable to convert to integer
@Alfredo: wouldn't this be easier:... Jan Fajerski

11/29/2018

01:15 PM Bug #37467 (Duplicate): ceph-volume: RuntimeError: dictionary changed size during iteration
Duplicate of https://tracker.ceph.com/issues/37356 Jan Fajerski
12:19 PM Bug #37467: ceph-volume: RuntimeError: dictionary changed size during iteration
Alfredo Deza wrote:
> Is this a Python 3 environment?
Yes
Volker Theile
12:11 PM Bug #37467: ceph-volume: RuntimeError: dictionary changed size during iteration
Fixed by https://github.com/ceph/ceph/pull/25321 Volker Theile
12:10 PM Bug #37467: ceph-volume: RuntimeError: dictionary changed size during iteration
Possibly a duplicate of https://tracker.ceph.com/issues/37356 Alfredo Deza
12:10 PM Bug #37467: ceph-volume: RuntimeError: dictionary changed size during iteration
Is this a Python 3 environment? Alfredo Deza
12:09 PM Bug #37467 (Duplicate): ceph-volume: RuntimeError: dictionary changed size during iteration
When running the command 'ceph-volume lvm batch --report --filestore /dev/vdg' the following error occurs.... Volker Theile
12:57 PM Bug #37470: ceph-volume: Python3 - name 'raw_input' is not defined
Fixed by https://github.com/ceph/ceph/pull/25324 Volker Theile
12:56 PM Bug #37470 (Resolved): ceph-volume: Python3 - name 'raw_input' is not defined
When running 'ceph-volume lvm batch --filestore /dev/vdg' on a Python3 environment (with fixed issue #37467) the foll... Volker Theile
08:55 AM Bug #37442: ceph-volume: Unable to convert to integer
Injected your patch and tried it again, command finishes now without errors! Thanks for your help ;) Kilian Ries

11/28/2018

09:42 PM Bug #37449 (Resolved): zap: improve zapping to remove all partitions and all LVs, encrypted or not
In order to get https://bugzilla.redhat.com/show_bug.cgi?id=1644847 (zapping by OSD ID) we need to improve the zappin... Alfredo Deza
09:25 PM Bug #37442: ceph-volume: Unable to convert to integer
This looks like a valid issue. If you are comfortable editing on the fly, could you try the following patch?... Alfredo Deza
08:54 PM Bug #37442: ceph-volume: Unable to convert to integer
here are the last lines from the logfile you requested:... Kilian Ries
05:48 PM Bug #37442: ceph-volume: Unable to convert to integer
That looks unexpected for sure... could you re-run (or reproduce) with:
> CEPH_VOLUME_DEBUG=1
So that we can s...
Alfredo Deza
04:52 PM Bug #37442 (Closed): ceph-volume: Unable to convert to integer
Hi,
just tried to manually provision my ceph osds, but got an error:...
Kilian Ries
04:20 PM Bug #37409 (In Progress): tests: use two NVMe devices for batch operations
Alfredo Deza
04:16 PM Feature #26864 (Resolved): ceph-volume lvm batch should be idempotent
https://github.com/ceph/ceph/pull/24404 Andrew Schoen
02:20 PM Feature #37083: ceph-volume should report device_id as part of the device list
As per the discussion on 28th November, the device id implementation must remain in c++ code and in ceph-volume separ... Anonymous

11/27/2018

08:31 PM Bug #37411 (Resolved): zapping a device not that is a pv but with no vgs fails
When running `ceph-volume lvm zap --destroy` against a device that is a pv but has not been used for a vg fails.
<...
Andrew Schoen
03:09 PM Bug #37409 (Resolved): tests: use two NVMe devices for batch operations
LVM rounds sizing down, which causes some setups to fail to deploy with more than one NVMe when not specifying size w... Alfredo Deza

11/26/2018

01:16 PM Bug #37390 (Resolved): c-v inventory returns invalid JSON
print() uses single-quotes by default, which is invalid JSON. Sebastian Wagner
12:58 PM Cleanup #37389 (Resolved): introduce class hierachy for strategies to remove duplicate code
See $subject.
Jan Fajerski

11/22/2018

04:57 PM Feature #37086 (In Progress): Add several flags to ceph-volume lvm batch
I had a look into the batch subcommand. The current argument stucture (with data devices as positional arguments and ... Jan Fajerski

11/21/2018

02:18 PM Bug #37356: ceph-volume lvm batch broken on py3 environments
https://github.com/ceph/ceph/pull/25203 Jan Fajerski
02:14 PM Bug #37356 (Resolved): ceph-volume lvm batch broken on py3 environments
... Jan Fajerski
09:21 AM Feature #37083: ceph-volume should report device_id as part of the device list
Alfredo Deza wrote:
> The choice is to have these configuration-management tasks be done with a configuration manage...
Anonymous

11/20/2018

08:29 PM Feature #37087: add ceph-volume zap <osd-id> <osd-uuid> subcommand

> Does it make sense to add a general ceph-volume zap command? This could re-use the lvm zap command in case its an...
Alfredo Deza
03:32 PM Feature #37087: add ceph-volume zap <osd-id> <osd-uuid> subcommand
zap looks good for managing lvm based osds. I think we´ll need a subcommand to zap ceph-disk deployed OSDs as well th... Jan Fajerski

11/19/2018

05:58 PM Feature #37083: ceph-volume should report device_id as part of the device list
The choice is to have these configuration-management tasks be done with a configuration management tool, not in ceph-... Alfredo Deza
05:08 PM Feature #37083: ceph-volume should report device_id as part of the device list
@alfredo : We have no choice here. We need to list the devices _before_ Ceph runs. So c-v have to be resilient when c... Anonymous
02:57 PM Feature #37083: ceph-volume should report device_id as part of the device list
There is an assumption here that ceph-volume can be installed without any other ceph component. A lot of the internal... Alfredo Deza
01:35 PM Feature #37087: add ceph-volume zap <osd-id> <osd-uuid> subcommand
Alfredo Deza wrote:
> I understand that removing an OSD is something that was discussed, I am not clear why the orch...
Jan Fajerski

11/16/2018

02:05 PM Feature #37087: add ceph-volume zap <osd-id> <osd-uuid> subcommand
I understand that removing an OSD is something that was discussed, I am not clear why the orchestrator can't issue th... Alfredo Deza
01:30 PM Feature #37087: add ceph-volume zap <osd-id> <osd-uuid> subcommand
Alfredo Deza wrote:
> Why would removing an OSD be something that ceph-volume needs to do?
This was identified in...
Jan Fajerski
09:52 AM Feature #37083: ceph-volume should report device_id as part of the device list
Yes that would be better to get a single implementation to avoid mismatches.
I see two set of calls of get_device_...
Anonymous

11/15/2018

12:24 PM Bug #27062 (Resolved): GPT devices are excluded
mimic PR: https://github.com/ceph/ceph/pull/25103
luminous PR: https://github.com/ceph/ceph/pull/25104
Alfredo Deza

11/14/2018

08:02 PM Feature #36603: Allow loop device as a LVM backend
:-(
But the second method of checking if device is valid (using `stat` package from python) recognizes loop devices...
Mariusz Strzelecki
07:59 PM Feature #36603 (Closed): Allow loop device as a LVM backend
Closing this as it is possible to use loop devices (if the LV is created before hand). We are not going to add functi... Alfredo Deza
06:05 PM Bug #27062 (In Progress): GPT devices are excluded
https://github.com/ceph/ceph/pull/25098 Andrew Schoen
04:42 PM Feature #37083: ceph-volume should report device_id as part of the device list
I had a look into this and I have two discussion points:
The current c++ implementation uses the ID_SERIAL udev pr...
Jan Fajerski
03:48 PM Feature #37083: ceph-volume should report device_id as part of the device list
It is not clear to me what spec I should adhere to. It seems like it isn't 1 piece of information added, but to compo... Alfredo Deza
12:21 PM Feature #37083: ceph-volume should report device_id as part of the device list
We need to share the same format so the manager can 'recognize' as disk c-v exposes as the same disk as the one it ha... Anonymous

11/13/2018

07:00 PM Bug #36768 (Resolved): tests failures when /dev/sda doesn't exist
mimic PR: https://github.com/ceph/ceph/pull/25066
luminous PR: https://github.com/ceph/ceph/pull/25067
Alfredo Deza
04:03 PM Feature #37087: add ceph-volume zap <osd-id> <osd-uuid> subcommand
Why would removing an OSD be something that ceph-volume needs to do?
There is already a ticket for destroying/zapp...
Alfredo Deza
02:37 PM Feature #37087 (New): add ceph-volume zap <osd-id> <osd-uuid> subcommand
Add subcommand to remove an OSD and its underlying storage. This should work for ceph-disk osds as well as lvm osds.
...
Jan Fajerski
04:01 PM Feature #37086: Add several flags to ceph-volume lvm batch
--dmcrypt, --objectstore, and --osds-per-device are already supported
I don't think that --num-slots or --shared-u...
Alfredo Deza
02:33 PM Feature #37086 (Resolved): Add several flags to ceph-volume lvm batch
as discussed in several orchestrator meetings, the batch subcommand needs several arg additions.
-replace-osd-ids=...
Jan Fajerski
03:56 PM Feature #37083: ceph-volume should report device_id as part of the device list
Seems like by 'ID' you mean something that gets composed somehow if certain information is not available. Do you need... Alfredo Deza
03:32 PM Feature #37083: ceph-volume should report device_id as part of the device list
As we can't rely on a running ceph cluster we'll sadly have to duplicate the code for this feature.
Code is here :
...
Anonymous
02:47 PM Feature #37083: ceph-volume should report device_id as part of the device list
Yeah, that's starting nautilus. A typical output looks like :
$ ceph device ls
DEVICE ...
Anonymous
12:03 PM Feature #37083: ceph-volume should report device_id as part of the device list
Could you expand a bit on this ticket? I've never seen `ceph device` before. Is that Nautilus only? When you say "lis... Alfredo Deza
11:08 AM Feature #37083 (Resolved): ceph-volume should report device_id as part of the device list
When listing the device list, the orchestrator needs to get the device_id as defined in "ceph device" command.
Anonymous
10:54 AM Feature #37082 (Rejected): Split ceph-volume into a separate package
As per the orchestrator sandwich, orchestrator should be in a position to list a host device list _before_ ceph is in... Anonymous

11/12/2018

06:00 PM Bug #36768 (Fix Under Review): tests failures when /dev/sda doesn't exist
master PR https://github.com/ceph/ceph/pull/25063 Alfredo Deza
04:21 PM Bug #36768 (Resolved): tests failures when /dev/sda doesn't exist
PR https://github.com/ceph/ceph/pull/24859 introduced new unit tests that rely implicitly on systems that have device... Alfredo Deza
 

Also available in: Atom