On 15.03.2022 10:10, Jimmy Spets wrote:
Thanks for your reply.
I have two things that I am unsure of:
- Is the OSD UUID the same for all OSD:s or should it be unique for
each?
It's unique and generated when you run ceph-volume lvm prepare or add an
OSD.
You can find OSD UUID/FSID for existi
Hi,
We had to recreate our Ceph cluster and it seems some legacy data was left
over. I think this is causing our valid OSD's to hang for 15-20 minutes
before starting up on a machine reboot.
When checking /var/log/ceph/ceph-volume.log I can see the following -
[2022-03-08 09:32:10,581][ceph_volu
Hi Simone,
There is a step that I see missing here - have you created a role? For
creating a role, you need to attach 'roles' caps to the user that you
created.
Also, what tool have you used to make the AssumeRoleWithWebIdentity call?
An example using boto3 is outlined in the documentation here:
h
Please correct the trust policy with the condition element that I pointed
out before. Also, Can you please try using AWS tools - boto3 or AWS STS
apis to make the AssumeRoleWithWebIdentity call. You can check the RGW log
files to see whether the call reaches RGW with the curl command.
Thanks,
Prit
This is now resolved. I simply found the old systemd files
inside /etc/systemd/system/multi-user.target.wants and disabled them which
automatically cleaned them up.
Thanks!
On Wed, 16 Mar 2022 at 09:30, Chris Page wrote:
> Hi,
>
> We had to recreate our Ceph cluster and it seems some legacy dat
Hi,
I'm having an issue on one of my nodes where all of it's OSD's take a long
time to come back online (between 10 and 15 minutes). In the Ceph log, it
sits on:
bluestore(/var/lib/ceph/osd/ceph-8) _open_db_and_around read-only:0 repair:0
Until eventually something changes which allows the start
Hi Chris,
could you please raise debug-bluestore level to e.g. 5 and share a new
startup log?
I presume it's fsck which is running at that point but let's double check.
And what Ceph version are we talking about?
Thanks,
Igor
On 3/16/2022 4:58 PM, Chris Page wrote:
Hi,
I'm having an is
Hi,
Has anyone figured whether those "lost" snaps are rediscoverable / trimmable?
All pgs in the cluster have been deep scrubbed since my previous email and I'm
not seeing any of that wasted space being recovered.
Regards,
LRT
-Original Message-
From: Dan van der Ster
To: technol...@li
>
> Thanks Igor,
So I stuck the debugging up to 5 and rebooted, and suddenly the OSD's are
coming back in no time again. Might this be because they were so recently
rebooted? I've added the log with debug below:
2022-03-16T14:31:30.031+ 7f739fd28f00 1
bluestore(/var/lib/ceph/osd/ceph-9) _m
Chris,
hmm... so I can't see any fsck output hence that's apparently not the
root cause.
Curios what happened to these OSDs and/or node before the first issue
appearance? Ceph upgrade ? Unexpected shutdown? Anything else notable?
Have you tried a single OSD restart multiple times and saw
Hi Igor,
> Curios what happened to these OSDs and/or node before the first issue
appearance? Ceph upgrade ? Unexpected shutdown? Anything else notable?
We originally had Ceph set up with some drives that we chose not to pursue.
On removing all of the drives things went a bit wayward and we ended
Hi all,
we have problem on our production cluster running nautilus (14.2.22).
Cluster is almost full and few month ago we noticed issues with slow peering - when we restart any osd (or host) it takes hours to finish
peering process, instead of minutes.
We noticed, that some pool contains 90k
The value of the 'aud' field in the token must be set in the Condition
element, checking it against 'app_id'. There is no need to add a custom
field 'app_id'.
The Ceph STS APIs have been tested using standard AWS tools (boto3 and
aws), so I'd suggest you to use them.
Thanks,
Pritha
On Wed, Mar 1
Interestingly this took a really long time (22 minutes overall).
However since the restart of OSD 10 will now restart no problem. I've put
the log below...
2022-03-16T15:26:09.181+ 7f0a2f26bf00 1
bluestore(/var/lib/ceph/osd/ceph-10) _mount path /var/lib/ceph/osd/ceph-10
2022-03-16T15:26:09.1
Hi Simone,
The condition element will be:
"StringEquals": {"mykeycloak.org.com/auth/realms/myrealm:app_id":"radosgw"}
Thanks,
Pritha
On Wed, Mar 16, 2022 at 9:44 PM wrote:
> Hi Pritha,
>
>
>
> I will test APIs with suggested tools.
>
>
>
> What is not clear to me is the aud and app_id.
>
>
>
Hi,
A few of the devs have been thinking about how we could make managing
multiple ceph clusters easier. At this point we're trying to
understand the requirements and problems that a multi-cluster feature
needs to fix, and need your help!
We've put together a short, 13 question survey;
https://fo
> We've put together a short, 13 question survey;
> https://forms.gle/E9cAx4f51Hq2FHQXA
>
FYI it is behind a sign-in wall
___
ceph-users mailing list -- ceph-users@ceph.io
To unsubscribe send an email to ceph-users-le...@ceph.io
Apologies - it was set to limit responses to 1 from an individual,
which uses a google account to prevent spamming the form.
I've turned that off now.
On Thu, Mar 17, 2022 at 11:21 AM Marc wrote:
>
> > We've put together a short, 13 question survey;
> > https://forms.gle/E9cAx4f51Hq2FHQXA
> >
>
Chris,
After you run "scrub repair" followed by a "scrub" without any issues, and
if the "damage ls" still shows you an error, try running "damage rm" and
re-run "scrub" to see if the system still reports a damage.
Please update the upstream tracker with your findings if possible.
--
Milind
On S
19 matches
Mail list logo
