meta data for this page
  •  

This is an old revision of the document!

Move DB to bigger storage

I.e: to move from 4GB partition to 30GB partition in case of spillover: 

1 OSD(s) experiencing BlueFS spillover

 osd.3 spilled over 985 MiB metadata from 'db' device (3.8 GiB used of 4.0 GiB) to slow device
lvdisplay
  --- Logical volume ---
  LV Path                /dev/ceph-66ad9fa6-0f65-4fc0-baeb-41089a509347/osd-block-4a9e142e-ed4d-4890-a018-5651c36071a5
  LV Name                osd-block-4a9e142e-ed4d-4890-a018-5651c36071a5

  --- Logical volume ---
  LV Path                /dev/ceph-ffd9911c-2d6c-43aa-b637-bec8a2092a30/osd-db-ad0a2108-bcca-4f15-9747-29d86624b911
  LV Name                osd-db-ad0a2108-bcca-4f15-9747-29d86624b911
  VG Name                ceph-ffd9911c-2d6c-43aa-b637-bec8a2092a30

  LV Size                4,00 GiB
  Current LE             1024
  
vgdisplay
    --- Volume group ---
  VG Name               ceph-ffd9911c-2d6c-43aa-b637-bec8a2092a30
  VG Size               4,00 GiB
  PE Size               4,00 MiB
  Total PE              1025
  Alloc PE / Size       1024 / 4,00 GiB
  Free  PE / Size       1 / 4,00 MiB

pvdisplay  
  --- Physical volume ---
  PV Name               /dev/nvme0n1p5
  VG Name               ceph-ffd9911c-2d6c-43aa-b637-bec8a2092a30
  PV Size               <4,01 GiB / not usable 2,00 MiB
  Allocatable           yes 
  PE Size               4,00 MiB
  Total PE              1025
  Free PE               1
  Allocated PE          1024
  PV UUID               pjYGgY-SS26-m8Sc-aWyQ-3zzY-f7Iz-XGoI90

resize exsiting DB partition

Remove separate DB and migrate data back to main storage: 

ceph osd set noout
systemctl stop ceph-osd@3.service
 
cat /var/lib/ceph/osd/ceph-3/fsid 
lvdisplay
 
ceph-volume lvm migrate --osd-id 3 --osd-fsid 024c05b3-6e22-4df1-b0af-5cb46725c5c8 --from db --target ceph-5582d170-f77e-495c-93c6-791d9310872c/osd-block-024c05b3-6e22-4df1-b0af-5cb46725c5c8
--> Migrate to existing, Source: ['--devs-source', '/var/lib/ceph/osd/ceph-0/block.db'] Target: /var/lib/ceph/osd/ceph-0/block
--> Migration successful.
 
systemctl start ceph-osd@3.service
ceph osd unset noout

create new one

Create new 41GB (41984MB) partition (/dev/nvme0n1p6) 

pvcreate /dev/nvme0n1p6
vgcreate ceph-db-40gb /dev/nvme0n1p6
lvcreate -n db-40gb -l 100%FREE ceph-db-40gb
ceph osd set noout
systemctl stop ceph-osd@3.service
 
cat /var/lib/ceph/osd/ceph-3/fsid 
 
ceph-volume lvm migrate --osd-id 3 --osd-fsid 4a9e142e-ed4d-4890-a018-5651c36071a5 --from db --target ceph-db-40gb/db-40gb
 
systemctl start ceph-osd@3.service
ceph osd unset noout

Move DB disks

In case of replace (full copy by dd) NVM drive where partition for OSD's DB is located, OSD service cannot start. Reason is that CEPH doesn't have permission to access device owned by root. To repair everything with one command: 

ceph-volume lvm activate --all

draft

$ systemctl stop ceph-osd@68
$ ceph-bluestore-tool --path /var/lib/ceph/osd/ceph-68 --devs-source
/var/lib/ceph/osd/ceph-68/block --dev-target
/var/lib/ceph/osd/ceph-68/block.db  bluefs-bdev-migrate
$ systemctl start ceph-osd@68