Forums

Home / Forums

You need to log in to create posts and topics. Login · Register

Can't add disk as cache on new node - old raid disks - best method to fix

I just added a 4th node to our petasan cluster.

I am now trying to add cache/journal/osd disks. The first disk i chose to add as a cache disk fails and I get the following in the log:

05/01/2021 11:18:16 ERROR Error executing : wipefs --all /dev/sdaa

05/01/2021 11:18:16 INFO Executing : wipefs --all /dev/sdaa

05/01/2021 11:18:15 INFO Start cleaning disk : sdaa

05/01/2021 11:18:10 INFO Start add cache job for disk sdaa.

05/01/2021 11:18:10 INFO -disk_name sdaa -partitions 4

05/01/2021 11:18:10 INFO params


When I log in at the console on the system and do a blkid I see that the disks used to be part of some raid sets:

/dev/sdab: UUID="e44dfd08-xxxxxxc8ca6adfc" UUID_SUB="b481dac4xxxxxx9c4aac7a3ac6" LABEL="ceph-osdxxxxxxu.edu:0" TYPE="linux_raid_member"

/dev/sdac: UUID="e44dfd08-xxxxxxc8ca6adfc" UUID_SUB="3e865b1xxxxxx6ad8842a4" LABEL="ceph-osdxxxxxxu.edu:0" TYPE="linux_raid_member"

/dev/sdad: UUID="e44dfd08-xxxxxx-284c8ca6adfc" UUID_SUB="fba56exxxxxx9e8bd2" LABEL="ceph-osdxxxxxxu.edu:0" TYPE="linux_raid_member"


lsblk reports this:

sdab                                                                                                   65:176  0 894.3G  0 disk  

└─md0                                                                                                   9:0    0  15.7T  0 raid6 

sdac                                                                                                   65:192  0 894.3G  0 disk  

└─md0                                                                                                   9:0    0  15.7T  0 raid6 

sdad                                                                                                   65:208  0 894.3G  0 disk  

└─md0                                                                                                   9:0    0  15.7T  0 raid6 

 


What is the best way to clear the old RAID info from these disks so I can add them into the cluster as Journal/Cache/OSD?

Thanks so much!

Neil

 

wipefs is failing, if you try the command manually, does it give more info on the error

wipefs --all /dev/sdaa

maybe the old RAID was activated, try to stop/delete it:

mdadm --stop /dev/md0
mdadm --remove /dev/md0

That was it. I stopped the old raids and then manually wiped the disks. I was then able to add them into as new members of the petasan volume.