Thank you for your help. I found it Steven On Mon, 29 Sept 2025 at 16:07, Pierre Riteau <[email protected]> wrote:
> Hi Steven, > > You have several ways to identify the drive from the OSD. > > Option 1: > ceph device ls | grep 'osd\.<id>$' > > This should return an output of the form: > <drive manufacturer>_<drive model>_<drive serial number> <host>:<dev name> > osd.<id> > > Option 2: > ceph osd metadata <id> > > This should show useful information in the "device_ids", "device_paths" > and "devices" keys fields. This will also include information about the > DB/WAL device, if in use. > > Then use the output of `lsblk -o NAME,SERIAL` or `lsblk -O` to identify > the drive from the serial number. > > On Mon, 29 Sept 2025 at 18:39, Steven Vacaroaia <[email protected]> wrote: > >> Hi, >> >> There must be something basic that I am missing but I have an "osd down" >> issue and cannot pinpoint which disk need to be replaced >> >> Steps I took >> >> 1. "Identify the problem" >> cepg osd stat >> 185 osds: 184 up (since 3d), 184 in (since 3d); epoch: e47876; 941 >> remapped >> pgs >> >> 2." identify the OSD" >> ceph osd tree | grep down >> 51 ssd_class 6.98630 osd.51 down 0 >> 1.00000 >> osd.51 ceph-host-3 >> stopped 3m ago 9w - 26.1G <unknown> >> <unknown> <unknown> >> >> 3. "check the block" >> ls -al /var/lib/ceph/0cfa836d-68b5-11f0-90bf-7cc2558e5ce8/osd.51 >> block -> >> >> /dev/ceph-419569a5-9a68-4688-8cf4-17ee450a8659/osd-block-75fe30c3-8cfa-49c7-80a1-6d79c3e4b4e6 >> >> 4. "find the device" >> ls -al >> >> /dev/ceph-419569a5-9a68-4688-8cf4-17ee450a8659/osd-block-75fe30c3-8cfa-49c7-80a1-6d79c3e4b4e6 >> /dev/dm-7 >> >> 5. "what type " >> blkid /dev/dm-7 >> /dev/dm-7: TYPE="ceph_bluestore" >> This is weird since osd 51 was type "ssd_class" so just a regular OSD >> deployed on SSD drive >> >> 5. which disk has dm-7 >> lsblk -o NAME,SERIAL | grep >> >> ceph-419569a5-9a68-4688-8cf4-17ee450a8659/osd-block-75fe30c3-8cfa-49c7-80a1-6d79c3e4b4e6 >> no results ?? >> >> >> blkid /dev/dm-7 >> /dev/dm-7: TYPE="ceph_bluestore" >> _______________________________________________ >> ceph-users mailing list -- [email protected] >> To unsubscribe send an email to [email protected] >> > _______________________________________________ ceph-users mailing list -- [email protected] To unsubscribe send an email to [email protected]
