boot fails on grub rescue

[giovvv]

I have just upgraded (apt-get upgrade) two machines to kernel 4.15.18-1-pve, then rebooted them. One, a Dell Poweredge R210 II, went flawlessly. The other, a Dell Poweredge R220, does not boot anymore, it ends in the grub rescue shell.

Both are configured with the disk controller set to passthrough (no hardware RAID), and have two 4TB disks with ZFS RAID1. Both have the same proxmox installation.

"ls" returns:
(proc) (hd0) (hd0,gpt9) (hd0,gpt2) (hd0,gpt1) (hd1) (hd1,gpt9) (hd1,gpt2) (hd1,gpt1)

"ls" on each of the above returns "no known filesystem detected", except on (hd0,gpt2) and (hd1,gpt2) which return a zfs type filesystem.

"set" returns, among other things:
cmdpath=(hd0)
prefix=(hd0,gpt9)/ROOT/pve-1@/boot/grub
root=hd0,gpt2

any help would be greatly appreciated.
Thanks

 

[dr.mythnick]

Hi,

i am hitting the same thing on 2 microserver gen8.

how did you resolve it ?

regards,
M

 

[giovvv]

I have no experience with hp microservers, but this is how I solved for my case:

1) no swap on ZFS (I don't remember if it was related to this particular failure, but still, don't do it; newer proxmox installers do not do it either).

2) no ZFS disk connected to any kind of pseudo-smart controller. Some are known to work (e.g. reflashed LSI cards) but Dell's RAID controller was not transparent enough for ZFS, even in passthrough mode. I disconnected the disks from that and connected directly to the motherboard's SATA ports. That worked.

I remember I could not salvage the crashed installation anyway, I had to rebuild everything from scratch.