Forwarding here in case anyone is seeing the same/similar issue, Amit gave really good pointers and a workaround :) Thanks Amit! On 11/25 16:08, Amit Ghadge wrote:

Yes, and if you want avoid in future update this flag to 0 by $echo 0 > /sys/block/sdx/queue/rotational

Thanks

On Wed, Nov 25, 2020 at 4:03 PM David Caro <dcaro@wikimedia.org> wrote:

...

Yep, you are right:

``` # cat /sys/block/sdd/queue/rotational 1 ```

I was looking to the code too but you got there before me :)

https://github.com/ceph/ceph/blob/25ac1528419371686740412616145703810a561f/s...

It might be an issue with the driver then reporting the wrong data. I'll look into it.

Do you mind if I reply on the list with this info? (or if you want you reply) I think this might help others too (and myself in the future xd)

Thanks Amit!

...

This might happen when the disk default sets 1 in /sys/block/sdx/queue/rotational , 1 for HDD and 0 for SSD, But we not see any problem till now.

-AmitG

On Wed, Nov 25, 2020 at 3:08 PM David Caro <dcaro@wikimedia.org> wrote:

...

Hi!

I have a nautilus ceph cluster, and today I restarted one of the osd daemons and spend some time trying to debug an error I was seeing in the log, though it seems the osd is actually working.

The error I was seeing is: ``` Nov 25 09:07:43 osd15 systemd[1]: Starting Ceph object storage daemon osd.44... Nov 25 09:07:43 osd15 systemd[1]: Started Ceph object storage daemon osd.44. Nov 25 09:07:47 osd15 ceph-osd[12230]: 2020-11-25 09:07:47.846 7f55395fbc80 -1 osd.44 106947 log_to_monitors {default=true} Nov 25 09:07:47 osd15 ceph-osd[12230]: 2020-11-25 09:07:47.850 7f55395fbc80 -1 osd.44 106947 mon_cmd_maybe_osd_create fail: 'osd.44

has

...

already bound to class 'ssd', can not reset class to 'hdd'; use 'ceph osd crush rm-device-class <id>' to remove old class first': (16) Device or resource busy ```

There's no other messages in the journal so at first I thought that

...

...

failed to start. But it seems to be up and working correctly anyhow.

There's no "hdd" class in my crush map: ``` # ceph osd crush class ls [ "ssd" ] ```

And that osd is actually of the correct class: ``` # ceph osd crush get-device-class osd.44 ssd ```

``` # uname -a Linux osd15 4.19.0-9-amd64 #1 SMP Debian 4.19.118-2+deb10u1 (2020-06-07) x86_64 GNU/Linux

# ceph --version ceph version 14.2.5-1-g23e76c7aa6 (23e76c7aa6e15817ffb6741aafbc95ca99f24cbb) nautilus (stable) ```

The osd shows up in the cluster and it's receiving load, so there seems to be no problem, but does anyone know what that error is about?

Thanks!

-- David Caro SRE - Cloud Services Wikimedia Foundation <https://wikimediafoundation.org/> PGP Signature: 7180 83A2 AC8B 314F B4CE 1171 4071 C7E1 D262 69C3

"Imagine a world in which every single human being can freely share in

On 11/25 15:50, Amit Ghadge wrote: the osd the

...

...

sum of all knowledge. That's our commitment." _______________________________________________ ceph-users mailing list -- ceph-users@ceph.io To unsubscribe send an email to ceph-users-leave@ceph.io

-- David Caro SRE - Cloud Services Wikimedia Foundation <https://wikimediafoundation.org/> PGP Signature: 7180 83A2 AC8B 314F B4CE 1171 4071 C7E1 D262 69C3

"Imagine a world in which every single human being can freely share in the sum of all knowledge. That's our commitment."

-- David Caro SRE - Cloud Services Wikimedia Foundation <https://wikimediafoundation.org/> PGP Signature: 7180 83A2 AC8B 314F B4CE 1171 4071 C7E1 D262 69C3 "Imagine a world in which every single human being can freely share in the sum of all knowledge. That's our commitment."