redhat-cip / puppet-ceph

Deploy Ceph using puppet

Home Page:http://ceph.com/

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

OSD disk reuse

opened this issue · comments

The first install of an OSD works. Subsequent installs work, but the OSD fails to see the disk as usable. Possibly some remnant data? I'm still researching the issue. It may be a bug in ceph itself, or perhaps we need to dev/zero the drive?

@dontalton did you try removing the data from the filesystem, then delete the partition, delete the partition table and eventually dd if=/dev/zero … the first sectors of the disk ?

I did much additional testing of this. Doing just a /dev/zero on the ENTIRE drive is enough to make the OSD reinstall correctly. The other steps don't seem necessary.

It turns out only a count of 100 is needed for this to work

dd if=/dev/zero of=/dev/DEVICE bs=1M count=100

OSD reloads work nice and smooth now. Before, the residual disk data would cause all kind of erratic behavior during the puppet run.