Re: ceph squid 19.2.2 - modifyng HDD OSD to use NVME DB
Thanks for the link and the details I have not installed cephadm and respectively ceph-volume on other hosts but the controller I am running Ubuntu 24.04.2 and noticed the versions available for both cephadm and ceph-volume are either 19.2.3 or 19.2.0 - no 19.2.2.?? On the controller, I have 19.2.2 Should I install 19.2.3 or 19.2.0 ? Steven On Fri, 20 Feb 2026 at 08:29, Dario Graña via ceph-users <ceph-users@ceph.io> wrote:
I used the approach recommended by Eugen a few months ago and it worked well.
I made a small change to the lv names, instead of osd.0 in # Create Logical Volume ceph:~ # lvcreate -L 5G -n ceph-osd0-db ceph-db
I used ceph:~ # lvcreate -L 5G -n ceph-db-${OSD_FSID} ceph-db
where ${OSD_FSID} is the following # Get the OSD's FSID [ceph: root@ceph /]# OSD_FSID=$(ceph-volume lvm list 0 | awk '/osd fsid/ {print $3}') fb69ba54-4d56-4c90-a855-6b350d186df5
On Thu, Feb 19, 2026 at 8:35 PM Eugen Block via ceph-users < ceph-users@ceph.io> wrote:
Maybe this helps:
https://heiterbiswolkig.blogs.nde.ag/2025/02/05/cephadm-migrate-db-wal-to-ne...
Note that cephadm shell is always executed locally on the host, not remotely, there’s no ‚--host‘ parameter.
Zitat von Steven Vacaroaia via ceph-users <ceph-users@ceph.io>:
Hi,
I have a 7 nodes cluster deployed using cephadm with a mixt of HDD, SSD and NVME
I would like to reconfigure some of the HDD OSDs and put their DB on
NVME
There are only 6 x18 TB OSDs and there is a dedicate , "empty" 1.6 TB NVME installed for this purpose
The cluster is configured as EC 4+2 and is not under heavy use
I would appreciate if you could advise what is the best way of achieving that
As far as I was able to find, there are these 2 options
1. remove and redeploy OSDs 2. use ceph-volume
Option 1 is pretty straight forward but more time consuming
For Option2 though I was not able to find any clear instructions
The first thing I stumbled one was how to use ceph-volume in a cephadm managed cluster on a host that does not have it
( "cephadm shell --host HOSTNAME -- ceph-volume ls" is complaining about "unrecognized arguments --host")
Any help,links,suggestions will be greatly appreciated
Steven _______________________________________________ ceph-users mailing list -- ceph-users@ceph.io To unsubscribe send an email to ceph-users-leave@ceph.io
_______________________________________________ ceph-users mailing list -- ceph-users@ceph.io To unsubscribe send an email to ceph-users-leave@ceph.io
_______________________________________________ ceph-users mailing list -- ceph-users@ceph.io To unsubscribe send an email to ceph-users-leave@ceph.io
If your cluster is managed by cephadm, don't interfer with ceph-volume outside of a cephadm shell. You won't get the results you're hoping for. It doesn't really make a difference if you use cephadm version 19.2.3 or 19.2.0 for this operation. Zitat von Steven Vacaroaia <stef97@gmail.com>:
Thanks for the link and the details
I have not installed cephadm and respectively ceph-volume on other hosts but the controller
I am running Ubuntu 24.04.2 and noticed the versions available for both cephadm and ceph-volume are either 19.2.3 or 19.2.0 - no 19.2.2.?? On the controller, I have 19.2.2
Should I install 19.2.3 or 19.2.0 ?
Steven
On Fri, 20 Feb 2026 at 08:29, Dario Graña via ceph-users <ceph-users@ceph.io> wrote:
I used the approach recommended by Eugen a few months ago and it worked well.
I made a small change to the lv names, instead of osd.0 in # Create Logical Volume ceph:~ # lvcreate -L 5G -n ceph-osd0-db ceph-db
I used ceph:~ # lvcreate -L 5G -n ceph-db-${OSD_FSID} ceph-db
where ${OSD_FSID} is the following # Get the OSD's FSID [ceph: root@ceph /]# OSD_FSID=$(ceph-volume lvm list 0 | awk '/osd fsid/ {print $3}') fb69ba54-4d56-4c90-a855-6b350d186df5
On Thu, Feb 19, 2026 at 8:35 PM Eugen Block via ceph-users < ceph-users@ceph.io> wrote:
Maybe this helps:
https://heiterbiswolkig.blogs.nde.ag/2025/02/05/cephadm-migrate-db-wal-to-ne...
Note that cephadm shell is always executed locally on the host, not remotely, there’s no ‚--host‘ parameter.
Zitat von Steven Vacaroaia via ceph-users <ceph-users@ceph.io>:
Hi,
I have a 7 nodes cluster deployed using cephadm with a mixt of HDD, SSD and NVME
I would like to reconfigure some of the HDD OSDs and put their DB on
NVME
There are only 6 x18 TB OSDs and there is a dedicate , "empty" 1.6 TB NVME installed for this purpose
The cluster is configured as EC 4+2 and is not under heavy use
I would appreciate if you could advise what is the best way of achieving that
As far as I was able to find, there are these 2 options
1. remove and redeploy OSDs 2. use ceph-volume
Option 1 is pretty straight forward but more time consuming
For Option2 though I was not able to find any clear instructions
The first thing I stumbled one was how to use ceph-volume in a cephadm managed cluster on a host that does not have it
( "cephadm shell --host HOSTNAME -- ceph-volume ls" is complaining about "unrecognized arguments --host")
Any help,links,suggestions will be greatly appreciated
Steven _______________________________________________ ceph-users mailing list -- ceph-users@ceph.io To unsubscribe send an email to ceph-users-leave@ceph.io
_______________________________________________ ceph-users mailing list -- ceph-users@ceph.io To unsubscribe send an email to ceph-users-leave@ceph.io
_______________________________________________ ceph-users mailing list -- ceph-users@ceph.io To unsubscribe send an email to ceph-users-leave@ceph.io
participants (2)
-
Eugen Block
-
Steven Vacaroaia