HassOS data resizing timeout

[agners]

Fresh install on a Lenovo M90n-IoT with an i3, 4gb ram and a 128gb low-performance NVMe SSD. Computer had windows before on the drive. Booted Ubuntu Live, and used Disk to restore and image (HAOS 12.2) as per HA tutorial. After 1m30s, timeout to resize the data partition: [DEPEND] Dependency failed for HassOS data resizing.

Command line (From CTRL-ALT-F2) with blkid showed part p8 to be correctly labeled (hassos-data) Ping to 8.8.8.8 and local network worked.

What worked: Booted Ubuntu Live, used Disk to FORMAT DISK, with option "empty no partitions" and "slow write zeroes everywhere". Then used Disk to restore an image (HAOS 12.2) as per HA tutorial.

On bootup, resizing the SWAP took 30s, resizing data was so fast I did not see going trough (so, less than 2s). ... and now it works.

I'm enclined to say it's a bug, as strictly following instructions gives a dead-end. Maybe the data resizing needs more time on a non-zeroed disk to perform the duty? Why put a timeout anyway, if the whole thing fails when it fails?? If zeroing the disk is critical, probably something important is incorrectly defined in the disk image? Like not zeroing a trailing sector after a partition table that can add garbage to it?

Anyway, that is my first welcome to HA, and it's already a bug hunt. Hope this helps other users and devs! :)

Originally posted by @LionRelaxe in https://github.com/home-assistant/operating-system/issues/2272#issuecomment-2077001018

[LionRelaxe]

If I may add, instead of making the timeout longer, it could be removed altogether. There is no point to "cancel after a certain amount of time" (a timeout), as the HA installation is not viable afterward. Better let it hang there, so the end-user is immediately aware of what and where the issue is. I would suggest to replace the countdown with: A "counting up" timer, with a remark in this vein: "Resizing Data partition for HassOS, expect a couple of minutes: 12m34s elapsed" Thanks!

[bryanvaz]

+1

This issue bogged me down for 4 hours because SSD was previously used as a VM storage drive. Did a dd if=/dev/zero then reimaged, and the new image immediately booted into the HA cli, no delay or timeout for the resizing step.

[BradB111]

+1

Trying to use a Zima board with an eMMC and this is the first time I've come across anything that reflects my experience, I have not been able to install HAOS yet however I'm mid way through a zero write.

I would like to echo the thoughts of @LionRelaxe. A timeout during a one time setup doesn't seem necessary.

UPDATE: 0 write worked