XBian DEVEL and STAGING repositories

[mk01]

thank for confirming.

that means also cifs mounts are ready when needed?

I'm aiming to completely remove the dependency to network for other services - but still make 100% sure, that no service / program will fail because of not available network share.

so xbian-net will be no service just task, trying to bring up eth0 and lo for cca 30s. then regardless of result will stop. this stop is still as dependency - but it can no longer happen that it will be stuck in started state and trying to mount shares (but failing due to typo in config, unavailability of share or upstream lib changes (the nfs4 vs nfs story) - ((or other bugs as for instance not considering CIFS by myself)).

when more feedback will be ok, I will move to staging and will write documentation as the packages there have many new functions implemented:
1. "bootloader" (bootmenu)
2. "telnet" kernel command line argument which makes RPI available via telnet starting immediately in INITRAMFS stage, then going through rootfs mount, pivot root during whole upstart phase until ssh and others are loaded. then it is stopped.so you never ever have to come to RPI personally (systems console is available too - so you actually see the real screen)
3. removed dependency on FSTYPE and on other (/etc configurations) entries which are relevant to boot media etc. simply system boots and runs from any media you copy it onto without need TO ADAPT OTHER PARTS
4. xbiancloner is now supporting creation of NFS ROOT including NFS tests (whether proper no_squash is used) and including automatic change in cmdline.txt. so you just click & reboot.
4a. supported (means tested) is also ext4 or f2fs.
1a. also booting from snapshots is supported - although this was tested only minimally (this applies for whole bootmenu).
5. new 3.12.x kernels with manually merged patches from BTRFS developers as I realized that Linus is months after official tree considered stable. Quite difference in locks and transaction management resulting in almost NO IOWAIT utilization - having amazing impact on perceived responsiveness. solved should be also years wrong "df" reporting and the absurd "filesystem full" problems.
6. support for such small details as for instance turn TV off during shutdown of RPI, run user script on shutdown
7. XBMC signals - I found a way how to solve the signal problems (XBMC QUIT, SHUTDOWN, REBOOT) - as result of internal python engine and addons with running blocked io ops and other unstoppable tasks - i'm fetching this states directly from XBMC and handed over to upstart directly.

and many more for sure.

[CurlyMo]

I disabled automount.

1a. It would be great to have the bootmenu updated after every apt update, so we can easily boot a previous snapshot after a failing apt.

Also consider releasing far more stables instead of just adding awesome features all the time without having reached a single stable to date.

---

[ 7.800168] Unable to handle kernel NULL pointer dereference at virtual address 00000a30
[ 7.800225] pgd = d62d4000
[ 7.800238] [00000a30] *pgd=00000000
[ 7.800264] Internal error: Oops: 5 [#1] ARM
[ 7.800278] Modules linked in: vchiq cifs

[mk01]

1a. bootmenu is dynamic. as it starts it scans all available .conf files on disks (partitions) + for btrfs it is listing all root/* snapshots as boot targets.
1b. each reboot triggered by updated system component is by default set to be booted WITH initramfs phase. this is there since beta2. if even BOOTMENU will be started on this occasion is only question of configuration parameter.

all other points are inseparable properties of solid, stable and failure proof user product. or you read it differently? then express yourself. or even better post patches (for whatever part they would be supposed to).

[f1vefour]

I think now that the three repositories are up and functional there should be an XBian 1.x without beta attached to the name that defaults to the stable repo.

Also simultaneously have a beta release that defaults to the staging repo.

In this way XBian can be both bleeding edge and stable depending on what the user wants.

[CurlyMo]

The latest devel kernel gives me a near immediate OOPS.

[Perkabalo]

(22nd Jan, 2014 03:04 AM)CurlyMo Wrote:  The latest devel kernel gives me a near immediate OOPS.

No issues for me so far, if you are referring to 1.3-6.10 - 3.12.7+

Ran upgrade with command: apt-get dist-upgrade

[CurlyMo]

Code:

Inst xbian-package-kernel [1.3-6.9a] (1.3-6.10 XBian devel repository:devel [armhf])
Conf xbian-package-kernel (1.3-6.10 XBian devel repository:devel [armhf])

---

About xbian-config
- The text is still far to wide (in the main menu).
- SSH text is aligned to the left.
- SSH enable gives me this:

Code:

/usr/local/include/xbian-config/modules/sshroot/function: line 40: /etc/init.d/ssh: No such file or directory

- Some of the dialog windows as far to high:
* xbian,root} password input window
* ssh root login success
* overlocking loading windows
* Hostname input and success window
* Kernel, only one installed message
* Licenses menu
* Services main window
- Video flags, last option "Disable the 'rainbow' screen at really start of RPI" --> "Disable pre-boot rainbow window"
- Does the kernel and resize sd still have any use?

[mk01]

(22nd Jan, 2014 03:58 AM)Perkabalo Wrote:  

(22nd Jan, 2014 03:04 AM)CurlyMo Wrote:  The latest devel kernel gives me a near immediate OOPS.

No issues for me so far, if you are referring to 1.3-6.10 - 3.12.7+

Ran upgrade with command: apt-get dist-upgrade

yes, that one. together with firmware 1.4.13-1.

[f1vefour]

My network is screwed, I don't know what this dummy0 network is either.

I boot up and my eth0 no longer gets an ip address.

I assign it an address (ifconfig eth0 10.10.10.228) and it works on the network but no Internet! I can't ping web addresses (yahoo.com) or ip addresses either (8.8.8.8).

I didn't even run an update, this just happened out of the blue.

[CurlyMo]

Are your DNS settings correct in /etc/resolv.conf?

[freem@n]

even with a wrong DNS you should at be able to ping IP adresses.... problem is obviously pretty bad

[f1vefour]

I will check resolv but I should be able to ping ip addresses. My network is fine as am connected to the same router with my phone and it has Internet and network on Pi is fine (after assigning IP) as I'm connected over SSH to the Pi.

I have never seen an issue like this.

[CurlyMo]

Without a proper resolv you can ping local ip address but just not external ones.

[f1vefour]

Thanks CurlyMo, will investigate when I get home.

I just wish I knew what happened.

[koper89]

I'm in for both devel and staging