Today I was replacing one of eight 8TB drives on my proxmox ve server ZFS pool. I needed to swap /dev/sdd, and long story short, while comparing the GPT table with /dev/sda, I deleted the GPT table on /dev/sda... instead of /dev/sdd the new replacement drive.
I got the great error "disk is currently in use"... dang it and realized I hit it on the wrong drive! But the server's still running until I reboot or reload I guess(!)
Anyway - my question is this: I have the partition table from before the screwup, and in fact all 8 of the drives have the same partition table, isn't there a way to recreate the GPT table and PRESTO - I'm fine?
Otherwise, I'll just let /dev/sdd finish resilvering and then replace /dev/sda...
Was really hoping someone here would help me save myself...
FWIW it's Proxmox 8.4.1 on a DL730, and I used the proxmox ve installer to build the 8 drive RAIDZ2 array as the boot and main store.
not that anybody cares, but I usually use these steps to replace a bad drive
sgdisk /dev/sda -R /dev/sdd (assuming /dev/sda is good and /dev/sdd is the new drive)
sgdisk -G /dev/sdd
proxmox-boot-tool format /dev/sdd2
proxmox-boot-tool init /dev/sdd2
lsblk -dpo NAME,SERIAL,WWN (to get the WWN)
zpool replace rpool /dev/disk/by-id/(bad disk id-disk3) /dev/disk/by-id/(newdisk id-disk3)
(and the reason I was dorking around in fdisk was the new drive was 512e not 4Kn
and I got an error because partition2 was only 1/8th the needed size because the block size was 512)
That's another thing I need help working around if it's possible...but for now using another 4Kn to replace the bad drive...
But can I do sgdisk /dev/sdb -R /dev/sda and proxmox-boot-tool format and init to fix my mistake? or is there something else I have to do?
I got the great error "disk is currently in use"... dang it and realized I hit it on the wrong drive! But the server's still running until I reboot or reload I guess(!)
Anyway - my question is this: I have the partition table from before the screwup, and in fact all 8 of the drives have the same partition table, isn't there a way to recreate the GPT table and PRESTO - I'm fine?
Otherwise, I'll just let /dev/sdd finish resilvering and then replace /dev/sda...
Was really hoping someone here would help me save myself...
FWIW it's Proxmox 8.4.1 on a DL730, and I used the proxmox ve installer to build the 8 drive RAIDZ2 array as the boot and main store.
not that anybody cares, but I usually use these steps to replace a bad drive
sgdisk /dev/sda -R /dev/sdd (assuming /dev/sda is good and /dev/sdd is the new drive)
sgdisk -G /dev/sdd
proxmox-boot-tool format /dev/sdd2
proxmox-boot-tool init /dev/sdd2
lsblk -dpo NAME,SERIAL,WWN (to get the WWN)
zpool replace rpool /dev/disk/by-id/(bad disk id-disk3) /dev/disk/by-id/(newdisk id-disk3)
(and the reason I was dorking around in fdisk was the new drive was 512e not 4Kn
and I got an error because partition2 was only 1/8th the needed size because the block size was 512)
That's another thing I need help working around if it's possible...but for now using another 4Kn to replace the bad drive...
But can I do sgdisk /dev/sdb -R /dev/sda and proxmox-boot-tool format and init to fix my mistake? or is there something else I have to do?
