In the beginning, I’m not quite a linux-guru, but I know the basics and I’m patient. But now I don’t know what to do next (after many attempts)… I bought a Jetson Nano to rebuild a home NVR in cheap way. I managed to migrate SD to USB SSD (I guess source was JetsonHacks how-to). USB boot works fine.
my extlinux.conf is:
LABEL primary
MENU LABEL primary kernel
LINUX /boot/Image
INITRD /boot/initrd
APPEND ${cbootargs} root=PARTUUID=f9c93f79-b453-49b7-9d64-b037d6226a94 rootwait rootfstype=ext4 console=ttyS0,115200n8 console=tty0 fbcon=map:0 net.ifnames=0
Now I need to connect another USB / NLSAS HDD for recordings.
So I connected this drive, find PARTUUID by blkid and edit fstab as:
PARTUUID=eb8ed98f-aacb-1b4f-97cf-d54a23f2c091 /mnt/big ext4 defaults 0 0
after the restart it does not boot and I have to disconnect the disk, then the boot passes … can anyone advise me what I’m doing wrong?
Is it possible to connect a second USB drive in boot-from-usb mode?
I can’t say for sure, but your fstab is mandating the UUID mount at that point, or boot will block and fail. I’d also suggest its file fsck check be after the main drive (increment from “0” to “1” for the non-rootfs partition). Possibly this will work (I’m also naming UUID instead of PARTUUID, not sure of the effect):
UUID=eb8ed98f-aacb-1b4f-97cf-d54a23f2c091 /mnt/big ext4 rw,suid,dev,exec,auto,nouser,async,nofail 0 1
…notice in that line “defaults” is replaced by a comma-delimited list of what “defaults” is an alias for, with the exception that I’ve also added “nofail”. The “nofail” implies the system will continue to boot if the partition cannot be found or mounted. If this were a mandatory mount for the system to work, then you would not use this; however, for an optional disk (especially an external optional disk which might not be present if unplugged) I consider this basic sanity.
I’m not sure if the PARTUUID would differ from UUID. I go by the UUID reported in “lsblk” as to what I use.
Same situation was with UUID. I’ll try again with UUID and rest you suggesting( /mnt/big ext4 rw,suid,dev,exec,auto,nouser,async,nofail 0 1)… but after few hours, now I’m on opencv building
If it still fails, then make sure (A) the mount point exists, and (B) the UUID matches what you see from lsblk.
what about UUID capitalization ? I saw a mention somewhere. Is it important?
I’m just worried that the problem will be at a different level than FSTAB. It will be related to the selection of - where to get kernel…
When one disk is connected, the jetson boot displays the “no sd” icon and continues to boot from the USB disk. This will not appear with two discs.
The “UUID=...” must have the “UUID” capitalized, but I don’t think it matters for the actual hexadecimal code. In my fstab files all of the actual partition UUIDs are lower case, but I think upper case will also work.
Is it not correct that you want this to boot from the USB first disk, and then after boot mount the second USB disk partition to “/mnt/big”? Not having the SD might matter for some issues, but if it is booting, then it shouldn’t matter. One possible “snafu” is if you are editing the wrong “/etc/fstab”. There might be one on some other rootfs if it isn’t booting the one you think it is booting.
Found solution. I depends on USB port priority, somehow… original/boot HDD must be connected to port before added HDD. So addon drive is in bottom usb close to LAN and boot drive is in top USB close to LAN port,