Dear Igor,
Is the ceph-volume lvm migrate command smart enough in octopus 15.2.14 to be
able to remove the db (included the wall) from the nvme even if it is
spilledover? I can’t compact back to normal many disk to not show spillover
warning.
I think Christian has the truth of the issue, my
I don't have much experience in recovering struggling EC pools
unfortunately. Looks like it can't find OSDs for 2 out of the 6
shards. Since you run EC 4+2 the data isn't lost but not 100% sure how
to make it healthy.
There was a thread a while back that had some similar issue albeit
possibly
Hi Istvan,
yeah both db and wal to slow migration are supported. And spillover
state isn't a show stopper for that.
On 10/2/2021 1:16 AM, Szabo, Istvan (Agoda) wrote:
Dear Igor,
Is the ceph-volume lvm migrate command smart enough in octopus 15.2.14
to be able to remove the db (included
In my setup I've disabled the sharding and preshard each bucket which needs
more then 1.1 millions of objects.
I don't think it's possible to cleanup, even if you run the command with the
really-really mean it, it will not do anything, I've tried already.
Istvan Szabo
Senior Infrastructure
On Fri, Oct 1, 2021 at 11:55 AM Andrew Gunnerson
wrote:
>
> Thanks for the info. Do shadow roots affect that calculation at all?
They're not supposed to, but I haven't worked in this code...but the
different names you get (with the ~ssd and ~hdd postfix) would
indicate not.
>
> In the regular
I just left it and I stopped using synchronous multisite replication. I'm only
using directional for a while which is working properly.
Istvan Szabo
Senior Infrastructure Engineer
---
Agoda Services Co., Ltd.
e: istvan.sz...@agoda.com
Thanks for the info. Do shadow roots affect that calculation at all?
In the regular "ceph osd crush tree", I don't see any cycles and it doesn't seem
like there are the same buckets under two roots (I only have one root):
root=default
rack=rack1
host=ssd-1
Hi Peter,
> When I check for circles I found that running the upmap balancer alone never
> seems to create
> any kind of circle in the graph
By a circle, do you mean something like this?
pg 1.a: 1->2 (upmap to put a chunk on 2 instead of 1)
pg 1.b: 2->3
pg 1.c: 3->1
If so, then it's not
It generally means, in CS terms, that you have a graph rather than a tree.
In other words, you have two roots, or other crush buckets, which
contain some of the same buckets/items underneath themselves.
On Fri, Oct 1, 2021 at 9:43 AM Harry G. Coin wrote:
>
> I asked as well, it seems nobody on
Hello, you are right, I must explain better
and I have to be more patient.
I have to install new clusters so I think I will use the last ceph version.
As far as object storage replication (if I understood well) there aren't
problems with 3 separated cluster.
As far as rbd-mirroring (in my case
I asked as well, it seems nobody on the list knows so far.
On 9/30/21 10:34 AM, Andrew Gunnerson wrote:
Hello,
I'm trying to figure out what overlapping roots entails with the default
scale-down autoscaling profile in Ceph Pacific. My test setup involves a CRUSH
map that looks like:
Ignazio;
If your first attempt at asking a question results in no responses, you might
consider why, before reposting.
I don't use RBD mirroring, so I can only supply theoretical information.
Googling RBD mirroring (for me) results in the below as the first result:
Now the rebalance started to continue but I always have 2 pg which is in
degraded state, what is weird that the up and acting osds are totally different
:/
PG_STAT STATE UP
UP_PRIMARY ACTING
On 01/10/2021 17:00, Szabo, Istvan (Agoda) wrote:
I just left it and I stopped using synchronous multisite replication. I'm only
using directional for a while which is working properly.
So you did setup a sync policy to only sync in one direction?
In my setup the secondary site does not
Marked as unfound, 2 of them shadow file, 1 is multipart trash, 1 was a normal
file but have copy, to be honest when I touch the min_size it will degrade pg
for a short period of time and not sure what it would cause in a not healthy
state. Let's see how far I can go.
Istvan Szabo
Senior
Hi,
does someone got a quick fix for falling behin shards in the metadata sync?
I can do a radosgw-admin metadata sync init and restart the rgw daemons to
get a full sync, but after a day the first shards falls behind, and after
two days I also get the message with "oldest incremental change not
Hey Istvan,
On 05/02/2021 03:00, Szabo, Istvan (Agoda) wrote:
I found 6-700 stale instances with the reshard stale instances list command.
Is there a way to clean it up (or actually should I clean it up)?
The stale instance rm doesn't work in multisite.
I observe a similar issue with some
3x SSD osd /nvme
Istvan Szabo
Senior Infrastructure Engineer
---
Agoda Services Co., Ltd.
e: istvan.sz...@agoda.com
---
-Original Message-
From: Igor Fedotov
Sent: Friday, October 1, 2021
Am 27.09.21 um 22:38 schrieb Josh Baergen:
>> I have a question regarding the last step. It seems to me that the ceph
>> balancer is not able to remove the upmaps
>> created by pgremapper, but instead creates new upmaps to balance the pgs
>> among osds.
> The balancer will prefer to remove
Hi,
We need to use rrdtool on a cephfs mount. But we get this error "could not
lock RRD".
Could you help me ?
Thanks
'Joffrey
___
ceph-users mailing list -- ceph-users@ceph.io
To unsubscribe send an email to ceph-users-le...@ceph.io
here a copy of pastbin. so there are no similar issues out there in the deep
internet and so i cant say the root cause:
Sep 30 16:24:47 pve04 systemd[1]: Started Ceph metadata server daemon.
Sep 30 16:24:47 pve04 ceph-mds[331479]: starting mds.pve04 at
Sep 30 16:24:47 pve04 ceph-mds[331479]:
On 10/1/21 07:07, von Hoesslin, Volker wrote:
hi!
my cephfs is broken and i can not recover the mds-daemons. yesterday i have update my ceph-cluster from v15 to v16 and i thought all working fine.
Do you use cephadm? There was (is?) an issue with the way cephadm
upgrades the MDS, see the
I can't access the pastebin, did you verify if you hit the same issue
as Stefan referenced
(https://lists.ceph.io/hyperkitty/list/ceph-users@ceph.io/thread/KQ5A5OWRIUEOJBC7VILBGDIKPQGJQIWN/)? Before deleting or rebuilding anything I would first check what the root cause is. As Stefan said,
Hi Ignazio,
On 10/1/21 14:16, Ignazio Cassano wrote:
I meant :
Cluster B pool name pool-clutser-B mirrored on Cluster A
Cluster C pool name pool-clutser-C mirrored on Cluster A
To me this should be feasible. As I said earlier, I never tested such a
setup though.
So on cluster A I should
Hello Arthur,
I meant :
Cluster B pool name pool-clutser-B mirrored on Cluster A
Cluster C pool name pool-clutser-C mirrored on Cluster A
So on cluster A I should have two rbd-mirror daemons
Ignazio
Il giorno ven 1 ott 2021 alle ore 13:35 Arthur Outhenin-Chalandre <
is there any chance to fix this? there are some "advanced metadata repair
tools" (https://docs.ceph.com/en/latest/cephfs/disaster-recovery-experts/) but
i'm not realy sure is it the right way to handle this issue?
i have created an "backup" bevor any tries with this command:
Hi,
On 10/1/21 11:25, Eugen Block wrote:
I don't know for sure but I believe you can have only one rbd mirror
daemon per cluster. So you can either configure one-way or two-way
mirroring between two clusters. With your example the third cluster
would then require two mirror daemons which is not
I have my dashboards and I can see that the db nvmes are always running on 100%
utilization (you can monitor with iostat -x 1) and it generates all the time
iowaits which is between 1-3.
I’m using nvme in front of the ssds.
Istvan Szabo
Senior Infrastructure Engineer
Hi,
I'm not entirely sure if this really is the same issue here. One of
our customers also works with k8s in openstack and I saw similar
messages. We never investigated it, I don't know if the customer did,
but one thing they encountered was that k8s didn't properly clean up
no, aside the update i had not input this command:
ceph fs set allow_standby_replay false
just only the lines from the proxmox update path ->
https://pve.proxmox.com/wiki/Ceph_Octopus_to_Pacific#Upgrade_all_CephFS_MDS_daemons
Von: Stefan Kooman
Gesendet:
On 10/1/21 11:48, von Hoesslin, Volker wrote:
i can not see what i have done wrong, this are my update steps:
Have you set allow_standby_replay to false?
Have you checked there were no more standby-replay daemons (ceph fs dump)?
Gr. Stefan
___
i can not see what i have done wrong, this are my update steps:
https://pve.proxmox.com/wiki/Ceph_Octopus_to_Pacific#Upgrade_all_CephFS_MDS_daemons
Von: Stefan Kooman
Gesendet: Freitag, 1. Oktober 2021 11:00:10
An: von Hoesslin, Volker; ceph-users@ceph.io
Thanks very much Yuval for your confirmation.
I have set the signature to v2 on the v1 aws-cli?
[ansibleuser@ceprgw01 z_ejbg]$ python3 topic_with_endpoint.py mytopic
{'TopicArn': 'arn:aws:sns:poc:app1:mytopic', 'ResponseMetadata': {'RequestId':
And how many OSDs are per single NVMe do you have?
On 10/1/2021 9:55 AM, Szabo, Istvan (Agoda) wrote:
I have my dashboards and I can see that the db nvmes are always running on 100%
utilization (you can monitor with iostat -x 1) and it generates all the time
iowaits which is between 1-3.
Thanks
Ignazio
Il giorno ven 1 ott 2021 alle ore 11:26 Eugen Block ha
scritto:
> Hi,
>
> I don't know for sure but I believe you can have only one rbd mirror
> daemon per cluster. So you can either configure one-way or two-way
> mirroring between two clusters. With your example the third
Hi,
I don't know for sure but I believe you can have only one rbd mirror
daemon per cluster. So you can either configure one-way or two-way
mirroring between two clusters. With your example the third cluster
would then require two mirror daemons which is not possible AFAIK. I
can't tell
Hi Yuval,
Thanks, after using aws-cli v1, no signature related issue occurs. However, I
can still see the issue related to sns topic creation.
[ansibleuser@ceprgw02 z_ejbg]$ aws --endpoint-url
http://objects.dev.xx.xx.xx:80 sns create-topic --name=mytopic
--attributes='{"push-endpoint":
Hi,
I'm not sure if setting min_size to 4 would also fix the PGs, but the
client IO would probably be restored. Marking it as lost is the last
straw according to this list, luckily I haven't been in such a
situation yet. So give it a try with min_size = 4 but don't forget to
increase
I'm not sure if anything else could break, but since the OSD isn't
starting anyway... I guess you could delete osd.3 from ceph auth:
ceph auth del osd.3
And then recreate it with:
ceph auth get-or-create osd.3 mon 'allow profile osd' osd 'allow *'
mgr 'allow profile osd'
[osd.3]
Hello All,
Please I would like to know if it is possibile two clusters can mirror rbd
to a third cluster.
In other words I have 3 separated ceph cluster : A B C.
I would like cluster A and cluster B can mirror some pools on cluster C.
Is it possible ?
Thanks
40 matches
Mail list logo