Network boot with nfs external persistent

[CATMANMember]

Hello and thank you for this great resource.
First of all I want to apologize for my English. I believe your Russian isn’t better )
I’m using the latest devel ISO and network boot with tftp (FreeBSD server) and the NSF (Ubuntu server) for rootfs and external persistent.

At the moment I have three major problems:
1) At the system boot stage there is a delay of about 40 seconds, ending with the error message “could not open framebuffer. no such file or dirrectory”
2) external-persistent dirrective do not work quite the way I expected: Instead of using an image file casper-rw it creates a directory called “geexbox”. Is this correct?
3) The system can not properly shutdown. It hangs forever on “Unmounting filesystems”. It takes about 4-5 minutes to unmount “mount-persistent” and “mount-data”, and then process stops with the line “unmounting /media”

I’ve tried various combinations pxelinux.сfg. The current version is:

LABEL geexbox
	MENU LABEL Start GBX...
	MENU DEFAULT
	KERNEL vmlinuz
	APPEND initrd=initrd splash=silent vga=789 loglevel=3 boot=nfs nfsroot=serverip:/path-to/geexbox rw external-persistent

(if I replace the value of external-persistent to “external-persistent=casper-rw” persistence feature doesnt’t work at all)

nfs-server configuration
/path-to/geexbox/ gbx.local(rw,async,no_root_squash,no_all_squash,no_subtree_check)

I will be glad to any advice. Thank you in advance for your help

[tomlohaveMember]

Hello,

1) ignore this
2) yep, it’s correct, network mounts should be mounted using this directory and others like mount-persistent and mount-data,
3) the problem : when you want to power off, connman is killed, so all network mounts are unreachable … once your problem, you can by-pass this using udhcpc instead of connamn (look in /etc/network)

[Author]

Posts