Safe mode or single user mode equivalent on SCALE?

Like it says. How do I boot in a way that doesnt mount pools, start services etc, using Scale?

I guess you could mess with GRUB. How well versed are you with that ? What is the reason to do that ?

I’d simply turn everything off manually and then turn stuff back on one by one to diagnose whatever you are diagnosing.

For example, encrypted pools usually require the passphrase to be entered on boot, so those won’t mount automagically. Then, if you have Apps running off said pool, they can’t start until the pool is decrypted, leaving just services to be enabled/disabled.

I suppose you could also isolate the NAS by doing all this work via the IPMI console, either in person or via the IPMI network interface while keeping the NAS network interface offline?

Keep in mind, some of the services, connections, etc. may require a reboot to ensure that they’re doing what they’re allegedly doing.

SCALE doesn’t have a dedicated safe-mode boot entry, so the closest equivalents are practical ones:

  1. Turn off what you’re diagnosing first. Disabling the services (and Apps) in the UI before rebooting keeps them disabled across reboots, so a normal reboot already gives you a minimal system you can re-enable one at a time.

  2. Encrypted pools won’t auto-import without the passphrase, which is handy if you want the data out of the way during maintenance. For unencrypted pools you can boot first, then keep the pool read-only (from the CLI: zpool import -o readonly=on ) before changing anything.

  3. If the UI itself misbehaves, the console/SSH login still works, so you can read journalctl there before touching config.

Farout’s GRUB point is fair too — an alternate boot entry or init target can approximate single-user mode — I’d just be cautious editing GRUB on a production NAS. Hope that helps!

The issue was troubleshooting an ambiguous issue of broken disk vs broken card vs disconnected disk vs idk what might also be.

So I wanted to boot in a way that wouldn’t touch the pool, and allow me to run diagnostic commands and troubleshooting commands secure in the knowledge that no miuddleware or other non-essential process was doing anything with any of the pool disks on or after boot, beyond enumeration and recognition, but also full zfs services were available for running import, export and zdb /status commands if they became needed.

On CORE as you say, Id have hit single user mode. I didnt know what to do on SCALE, and Im absolutely unfamiliar entirely with linux or grub. Just wanted CLI + standard ZFS + usual shell commands (e.g. lspci) usable.

I now know whats up causing the original issue. So my problem has shifted somewhat.

But Id still like the answer on this, as this info will surely be needed in future some time too.