is there a way to get the actual ashift value of a zpool? zpool get ashift rpool reports 0, but I assume it should be 12

ah, mdb -k and ::spa_config

so it's 9

zdb -C also will tell you.

ah right

thanks!

tsoome: I hope you received my mail. I forgot to mention that I tried it in this case with driver/graphics/nvidia-390 - but the error is similiar on all other versions.

seems not to be my week. mirrored new nvme ssd's to my rpool, removed the sata ssd, rebooted and now I can't boot. I booted from an OI install stick and ran bootadm install-bootload -R rpool, but still nothing. sigh.

Agnar: happy new bug

yeah, even from load command line, chain on any lsdev doesn't work: "error while reading the file"

the question is: can I make the sata ssd that I detached from the pool make an rpool again without loosing the data:

?

I'll try later to import the rpool via boot stick, mirror the sata ssd to it and then try to boot from the sata again.

Did you detach? Or did you "zpool split"?

detach unfortunately

If you detached (like from a mirror) I don't know the difficulty in bringing it back.

(it's been a while, and I always use `zpool split`)

I once had a repoducable bug with 3-way mirrors and zpool split in sol11.2 - since then I'm careful with split ;)

I should reconsider my default behaviour now :)

I've never split from anything more than 2-way. For all I know you might be able to reproduce that bug on illumos. (If you have the reproduction handy I have a VMware environment where I can test a 3-way mirror rather easily.)

rmustacc: hi, i was looking for execvpe() implementation and noticed that you have code.illumos.org/c/illumos-gate/+/2066; it seems it's not ready for review yet though?

→ CODE REVIEW 2066: 7125 Provide execvpe (NEW) | illumos.org/issues/7125

yuripv: I drafted it, but never had a chance to write tests or use it anywhere.

yuripv: What'd be most helpful to you? I guess I could refresh it and maybe force myself to write some tests?