Forums

Home / Forums

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

how to replace disks when using cache?

Hi, I had a failing disk in a node with 4 OSDs and 1 SSD acting as cache (4 partitions). I marked the failing disk as "out", I waited for rebalancing, and I stopped OSD, and replaced the disk.

However, the cache ssd marks 0/4 available partitions, and lvm shows:

......

/dev/sdc3  ps-2e1fa5e6-1e20-4a48-8390-493b59378c72-wc-osd.0  [cache]
[unknown]  ps-2e1fa5e6-1e20-4a48-8390-493b59378c72-wc-osd.0  [main_wcorig]
.......

Where OSD 0 was the removed one. How should we proceed to change failing disks when using cache?

(we are using petasan 2.7.2)

Thanks!

was the OSD deleted from ui or manually ?

I stopped the OSD from shell, and I removed from GUI.

Note, I ended up reinstalling the whole node and replacing it, now everything is fine. But I'd like to know the official procedure to replace a failing disk, behind a cache drive.

Thanks for your help!