[mk01]

@CurlyMo

this is referring to what variant? NFS root?

[CurlyMo]

Fairly clean beta 2 install (without NFS).

[mk01]

ok, just some notes - static ip configured by the router is still dhcp for xbian. if ip is configured before system starts networking, this IF needs to be configured as "inet manual" in interfaces file. otherwise debian skips interface as "already configured" -> what means IF set as auto is missing configuration -> what automatically makes static-network-up not met -> networking not met -> runlevels not started -> finally triggering failsafe boot dropping to shell.

then consider network mount in fstab not being mounted -> mountall job will not report events as network filesystems, filesystem, sometimes not even virtual filesystems.

---

I will put shortly updates (to devel) with bootmenu finally (it can currently boot any system, any partition. no snaps and nfs from bootmenu yet), telnet available during initramfs phase and immediately with init start - making possible to access RPI at all times remotely if needed (stops as soon as ssh is available). only eth0 (or other IF) is needed, no other dependency. take over the console is possible also remotely with actual console screen. also all hard dependency to btrfs is removed. i tried to boot XBian from ext4, f2fs, xfs - including updating, upgrading. then option to choose any FS for cloning function will be added (including NFS).

[CurlyMo]

Concretely, how can i currently fix the network boot issue?

[snowman58]

Re-Booting

I have a RPI based installation of the latest release (installed using the Windows based installer).

It works sometimes, but most of the time it goes into a re-boot loop;
"Starting xbmc"
blank screen for 10-20 seconds
Initial (in my case weather) screen for a few seconds
a command line screen for a second.

Loop starts over.

I have re-installed three times hoping for a "magic" solution!

Suggestions?

Thanks

[mk01]

(8th Jan, 2014 12:45 AM)CurlyMo Wrote:  Concretely, how can i currently fix the network boot issue?

made it bold in previous post. literally it means that configuration for eth0 must look like this:

Code:

auto eth0
iface eth0 inet manual

---

(8th Jan, 2014 02:59 AM)snowman58 Wrote:  I have a RPI based installation of the latest release (installed using the Windows based installer).

It works sometimes, but most of the time it goes into a re-boot loop;
"Starting xbmc"
blank screen for 10-20 seconds
Initial (in my case weather) screen for a few seconds
a command line screen for a second.

please check if this is your case:

https://github.com/xbianonpi/xbian/issues/478#issuecomment-29559179

[snowman58]

(8th Jan, 2014 02:59 AM)snowman58 Wrote:  I have a RPI based installation of the latest release (installed using the Windows based installer).

It works sometimes, but most of the time it goes into a re-boot loop;
"Starting xbmc"
blank screen for 10-20 seconds
Initial (in my case weather) screen for a few seconds
a command line screen for a second.

please check if this is your case:

https://github.com/xbianonpi/xbian/issues/478#issuecomment-29559179
[/quote]

Thanks for the help.

Not sure what issue the link is addressing, but it looks like the fix has already been incorporated in the latest release (which I am using). It may be a related issue, but I am not smart enough to see the relationship.

Thanks!

[CurlyMo]

When i have:

Code:

auto eth0 lo
iface eth0 inet manual

eth0 doesn't come up...

---

This works, but the network splash still doesn't resolve until it ends in the shell:

Code:

auto eth0 lo
iface lo inet loopback
iface eth0 inet dhcp

allow-hotplug wlan0
iface wlan0 inet manual
wpa-roam /etc/wpa_supplicant/wpa_supplicant.conf
wireless-power off
iface default inet dhcp

[mk01]

we are not talking about the same thing. with "Concretely, how can i currently fix the network boot issue?" I was expecting kernel auto configuration so eth0 is up, just needs to be marked as up for system and emit events.

post /run/upstart-ev.log to pastbin + dmesg.

---

oh now I know. i can't find the other post with your fstab but if there is smb share and no nfs shares, then this other post should be solution:

http://forum.xbian.org/thread-1970-post-19492.html#pid19492

[CurlyMo]

3 NFS shares + 1 samba share.

http://pastebin.com/hBmxD2aN
http://pastebin.com/vsRW9Xyu

[mk01]

nfs 2/3 or 4?

possibly this http://forum.xbian.org/thread-1770-post-18776.html#pid18776

if not send

Code:

tail /var/log/upstart/xbian-net.log

[CurlyMo]

What worked was removing one failing NFS mount

Why does it crash on this?

[mk01]

@CurlyMo

most of the setups with network shares entries in fstab have some dependency on it. you symlinks, me too for shared xbmc settings, others for db being on NAS etc. so the workflow:
1. net up
2. net mounts from etc as must
3. xbmc start
4. 100s failsafe trigger checking net and xbmc. if this is missing it opens console and force starts all nets (regardless of "auto" etc).

the "issue" lies in 300s timeout for step 2. no problem to split it like 20s net mount hard then other dependencies are allowed, mount can continue but will emit event as done

is xbmc is dependent on it it will fail as well but yes. it won't take 100s.

[CurlyMo]

Why not add a custom XBian fstab option called "required".

In my case:

Code:

//10.0.0.140/rpi         /media/rpi      cifs    username=xxxx,password=xxxx,rw,file_mode=0777,dir_mode=0777,nounix,noserver​ino,required 0 0
10.0.0.140:/audio      /media/audio     nfs     defaults 0 0
10.0.0.140:/movies     /media/movies    nfs     defaults 0 0
10.0.0.140:/tv        /media/tv       nfs     defaults 0 0
/dev/mmcblk0p1        /boot                   xbian   rw                              0       1
/dev/root             /home                   xbian   subvol=home/@,noatime           0       0
/dev/root             /lib/modules            xbian   subvol=modules/@,noatime        0       0
/dev/root             /                       xbian   noatime                         0       0
none            /run/lock                       tmpfs                   noauto                  0       0
none            /run/shm                        tmpfs                   noauto                  0       0
none            /run/user                       tmpfs                   noauto                  0       0
none            /sys/kernel/debug               debugfs                 noauto                  0       0
none            /sys/kernel/security            securityfs              noauto                  0       0

So, it will only fail when it's unable to mount my /media/rpi mountpoint and not when any of my nfs shares were unable to mount.

[mk01]

I would prefer existing defined common keywords.

since upstart mountall there is "nobootwait". Debian's linux-utils & libs is again too old not supporting it - meaning manual "mount /media/tv" run will throw error so "mount -s /media/tv" would be needed but it is supported by upstart processes.

actual devel xbian-update is putting it already to /home, /lib/modules and / - this makes no need to commend them out in case of booting from non-btrfs filesystem.