FAI install with huge JBODs present - disk/by-id?

Steffen Grunewald steffen.grunewald at aei.mpg.de
Tue Dec 17 10:07:56 CET 2024


Good morning,

after a very promising start, I experienced a set-back yesterday with an
otherwise proven-to-work FAI setup, and need an alternative now.

Over the past years, I wrote, used, and improved, a scheme that would
ignore disk_config completely and work with hooks, to create a ZFS root
pool (and possibly a data pool as well) over disks selected via patterns
(usually /dev/disk/by-id/ata-MANUFACTURER).

This initially also worked for my recent setup, a BeeGFS server with two
OS disks (ata-*), a handful NVMes and two times 60 disks over multipath,
in total 247 disks seen by the kernel.

After setting up multipath and storage zpools, I can no longer boot into
the system as long as at least one JBOD is connected - grub 2.06 gets an
"error: out of memory" and even refuses such elementary tasks as
"ls (...)/".
Upgrading to grub 2.12 (from bookworm-backports) makes things worse as
the screen stays dark, not even the greeting "Welcome to GRUB!" comes up.

I'm afraid I've got to go back to md raid, but with all disks connected,
how do I find the two system disks, if not via /dev/disk/by-id? (a quick
check with "sysinfo" shows them at /dev/sda and /dev/sddi)

Is there a regular way via disk_config, or do I have to invent a 17-sided
wheel myself?

Thanks for listening,
 S

-- 
Steffen Grunewald, Cluster Administrator
Max Planck Institute for Gravitational Physics (Albert Einstein Institute)
Am Mühlenberg 1 * D-14476 Potsdam-Golm * Germany
~~~
Fon: +49-331-567 7274
Mail: steffen.grunewald(at)aei.mpg.de
~~~


More information about the linux-fai mailing list