Hello! I have a problem with one node. After reboot it failed to start pve-manager and the services it said there is no free space left on /. So i managed to free about 300Mb removing unnecessary caches and packages but df said root is still 100% full. Then i took another drive installed clean PVE on it and imported my ZFS pool as second storage. I can see disk image of my vm now so i created VM and added this image as second drive. When i try to move it to local storage i get an error "dataset does not exist" however zfs mirror is online. Please help....
[SOLVED] Help needed in recovering node with ZFS mirror
- Thread starter Vasilij Lebedinskij
- Start date
-
- Tags
- proxmox 4.2 recover
Yes. I successfully added my pool to new node but i can't move disk image to another storage to free space and recover old node. What does error "dataset does not exist" mean?
Yes! But in my case have zfs storage and it refuses to move disk image from it on new node. On old node i have no space left so i can't start pve-manager and move disk image to second lvm storage. I'm in deadend...
Today i'm trying to expand ZFS pool so i can recover old node.
So now i have degraded pool with only /dev/sda2 and missing /dev/sdc2
I detached /dev/sdc drive and pluged drive with more capacity 320GB (/dev/sdb), cloned partition layout and now resilvering my pool. After that i want to add another drive with same capacity as new one (dev/sdc), clone partition layout on it, detach first drive (dev/sda) and resize partition with zfs on last one (dev/sdc), after resilver and detach second drive (dev/sdb). So i should have degraded pool with a lot of free space in the end.... i hope. Is this right way i'm going?
So now i have degraded pool with only /dev/sda2 and missing /dev/sdc2
I detached /dev/sdc drive and pluged drive with more capacity 320GB (/dev/sdb), cloned partition layout and now resilvering my pool. After that i want to add another drive with same capacity as new one (dev/sdc), clone partition layout on it, detach first drive (dev/sda) and resize partition with zfs on last one (dev/sdc), after resilver and detach second drive (dev/sdb). So i should have degraded pool with a lot of free space in the end.... i hope. Is this right way i'm going?
Your problem is not related to Proxmox, but more to ZFS specific…
https://pthree.org/2012/04/17/install-zfs-on-debian-gnulinux/
https://pthree.org/2012/04/17/install-zfs-on-debian-gnulinux/
This procedure should work. I did it a while ago, but I did not replace the small drive in the mirror and just left it as a single drive.
You will need to issue this command to tell ZFS to use the new full size of your drives after the small drive is removed:
Code:
zpool online -e rpool /dev/sda2 /dev/sdb2assuming you are left with those two devices as your pool mirror members.
Thanks for advise but i gone another way to solve my problem! I just added another mirror and upgraded my pool to raid 10. My pool is up again.