+- Forum (http://forum.xbian.org)
+-- Forum: Software (/forum-6.html)
+--- Forum: Testing & Experimental (/forum-21.html)
+--- Thread: beta2 (/thread-1312.html)
RE: beta2 - mk01 - 30th Sep, 2013 12:05 PM
I just wanted to see that despite chip 5730(5370) or whatever it is recognized as 3070, but then internally RFID is wrong, so the patch is correct. it actually allows 5370 driver do get 3070 ID. just compiling kernel.
log into ssh
Code:
cd /tmp
sudo -i
wget http://ivka57.dyndns-ip.com/others/xbian-package-kernel1.3-5.4.deb
dpkg -i xbian-package-kernel1.3-5.4.deb
rebootplease report
if you confirm, I will put into repo as part of Beta2
(30th Sep, 2013 03:32 AM)webjib Wrote: working perfectly now
perfect, thank for confirming this.
did you try the xbmc? you can full copy & paste the commands, it takes no more than 5 minutes. after you quit xbmc, the second block will rollback exact FS state from time before we moved over the old binaries.
you are really done in 5 min. ok, 7m.
RE: beta2 - darrylb - 30th Sep, 2013 02:00 PM
Well, looks like that fixed it... I ran those commands and joined it to my wireless network without any effort. Ill let you know how the reliability of the connection is over the next few days. I have my router auto-restart nightly, so will be curious to see if it recovers from that.
(30th Sep, 2013 12:05 PM)mk01 Wrote: I just wanted to see that despite chip 5730(5370) or whatever it is recognized as 3070, but then internally RFID is wrong, so the patch is correct. it actually allows 5370 driver do get 3070 ID. just compiling kernel.
log into ssh
Code:
cd /tmp
sudo -i
wget http://ivka57.dyndns-ip.com/others/xbian-package-kernel1.3-5.4.deb
dpkg -i xbian-package-kernel1.3-5.4.deb
reboot
please report
if you confirm, I will put into repo as part of Beta2
(30th Sep, 2013 03:32 AM)webjib Wrote: working perfectly now
perfect, thank for confirming this.
did you try the xbmc? you can full copy & paste the commands, it takes no more than 5 minutes. after you quit xbmc, the second block will rollback exact FS state from time before we moved over the old binaries.
you are really done in 5 min. ok, 7m.
RE: beta2 - mk01 - 30th Sep, 2013 03:23 PM
(30th Sep, 2013 02:00 PM)darrylb Wrote: Well, looks like that fixed it... I ran those commands and joined it to my wireless network without any effort. Ill let you know how the reliability of the connection is over the next few days. I have my router auto-restart nightly, so will be curious to see if it recovers from that.
found some comments about issues with dropouts with changing signal quality, but it was older driver without power output control.
everything can be different with you.
so I will put the patch into XBian kernel patch collection as needed and working, yes?
@wind-rider
nfs root:
I was possible to boot, but only without using initramfs.gz. The initramfs.gz was missing mount.nfs helper binary.
Changes for booting without initramfs.gz:
Code:
1) comment "initramfs.gz" stuff out in config.txt
2) in cmdline.txt change root to "root=/dev/nfs"
3) remove whole "rootflags=.*"
4) add "rw"
5) change "rootfstype=nfs"
6) add nfsroot=IP:/path/to/nfs,udp|tcp,vers=3
7) add ip=dhcp (asks kernel to configure dhcp on all network cards it sees during init
8) removed /home and /lib/modules from /etc/fstab
9) copied XBian before to prepared export (no_root_squash is essential there). done it with rsync in three steps:
a) rsync -avx --progress --delete / /mnt/nfs_root
b) rsync -avx --progress --delete /home/ /mnt/nfs_root/home/
c) rsync -avx --progress --delete /lib/modules/ /mnt/nfs_root/lib/modules/comment to point 6)
(udp OR tcp) depends on NAS capabilities and what you feel better in terms of speed, latency. also wsize= and rsize= are parameters to play with. but with nfsroot the idea is opposite to make buffer BIG for throughput, the smaller buffer the less latency (packets doesn't need to wait until buffer filled). I had best experiences with 2048-4096, but again depends. essential is correct IP:/path and despite nolock the only option to mount nfs<4 without tools, don't add it to options, kernel will not mount it then - without apparent error message. and if initramfs.gz will be used, I will add the nolock there if missing. also, vers=X is essential.
for initramfs.gz option make work, added nfsmount from klibc package and two conditions, other parts were not broken. just doing some edits to upstart, will issue then xbian-update 1.0.2-2 and xbian-package-initramfs-tools (whatever next version). xbmc starts even without the edits, but some of the scripts should not do what they do for network mounted roots.
you was probably stuck on initramfs error unable to mount root (due to the missing helper tool).
forgot quit important info: with kernel configuring IP and doing IFUP, /etc/network/interfaces has not contain "auto" for eth0 (if eth0 is the one you use) and instead of dhcp/static there must be manual. no-auto means system will not be waiting for eth0 appear up (because already is), will trigger "static-network-up" event as should - otherwise it won't be generated and manual means system will not touch net config if is configured
RE: beta2 - mk01 - 30th Sep, 2013 06:00 PM
(25th Sep, 2013 01:25 AM)IriDium Wrote: What is actually backed up in ssh "Backup / Home to file"?
I created a backup, then moved ~./*.zip to a folder Addons. Put Backup....gz into Put_here_to_restore and rebooted. Nothing had changed. I had expected the zip files to be back where they were when the image was created or am I missing something here?
you don't have to reboot. after you put the .gz file into the "drop box" (put_here_to_restore) folder all will happen automatically.
the atomic image will be tested for corruptions / if is ok, then /home will be replaced with it. after this is checked again, xbmc will be restarted.
no reboot needed, no intervention needed.
RE: beta2 - Airwolfuk - 30th Sep, 2013 07:44 PM
(30th Sep, 2013 02:36 AM)Airwolfuk Wrote:(29th Sep, 2013 05:49 PM)Airwolfuk Wrote: connection pages. I can only setup wifi using ssh Xbian Config and even then that has issues as previously reported. Once my wifi is setup it runs pretty good. With external power I get HD streaming over http most of the time.
what is your xbian-package-config-xbmc and -shell version ? can you start xbian-config XBMC and post few (10-15 lines) of python errors? does "/home/xbian/.xbmc/userdata/addon_data/plugin.xbianconfig" directory exists with those files?
"advancedmode confirmationonchange lastupdatecheck notifyonerror notifyonsuccess"
btw: I aligned the ssh result codes so if you update (at least) xbian-package-config-shell it should not be reverted anymore.
My versions were from last week (Monday). I am running the upgrade now and will report back tomorrow.
[/quote]
attached is what is in home/xbian/.xbmc/userdata/addon_data/plugin.xbianconfig
RE: beta2 - mk01 - 30th Sep, 2013 08:14 PM
your disk free space is fine and ownership (not the forced displayed via windows, but actual) is xbian?
la -la home/xbian/.xbmc/userdata/addon_data/plugin.xbianconfig
RE: beta2 - Airwolfuk - 30th Sep, 2013 08:21 PM
(30th Sep, 2013 08:14 PM)mk01 Wrote: your disk free space is fine and ownership (not the forced displayed via windows, but actual) is xbian?
la -la home/xbian/.xbmc/userdata/addon_data/plugin.xbianconfig
owner is Xbian
I just realized that there is no system tab in xbian-config which is why there are no network options. Setting advanced options is still not working even after updating/ upgrading this morning via ssh xbian-config.
RE: beta2 - mk01 - 30th Sep, 2013 08:46 PM
quit xbmc, delete whole folder /home/xbian/.xbmc/userdata/addon_data/plugin.xbianconfig
there should appear the first run welcome window and the settings should be recreated
RE: beta2 - webjib - 30th Sep, 2013 08:53 PM
(30th Sep, 2013 12:05 PM)mk01 Wrote: did you try the xbmc? you can full copy & paste the commands, it takes no more than 5 minutes. after you quit xbmc, the second block will rollback exact FS state from time before we moved over the old binaries.
you are really done in 5 min. ok, 7m.
I did not run the test, because I've tried another SD card with 1) xbian beta 1 and 2) Openelec
I was wrong, the problem is the same on all distributions, so I guess it's an issue in xbmc core. Not a major issue, but I'll investigate through xbmc forums...
RE: beta2 - Airwolfuk - 30th Sep, 2013 09:36 PM
(30th Sep, 2013 08:46 PM)mk01 Wrote: quit xbmc, delete whole folder /home/xbian/.xbmc/userdata/addon_data/plugin.xbianconfig
there should appear the first run welcome window and the settings should be recreated
Didn't work - still the same issues.
Update - I started over again with a fresh beta 1.1 - 1.2 install and can report that Xbian-Config is working as expected. Setting up my Wifi is now working (I can set up wifi from the GUI as well as ssh xbian-config), i can select advanced mode etc.
I don't know what happened or got corrupted on my 'old' install, but i am leaving that behind for my new one. :-)
Thanks for everyone's help and i'll let you know how i get on.
RE: beta2 - belese - 1st Oct, 2013 01:19 AM
Hi,
i finally managed to get a pi "working", i've no screen attached and very poor internet connection,
so xbmc itself is hard to test, i just use vnc to see interface.
and i've just find two small 'problems',
first is 'python service' is not recognise as running service by xbian-config, but the service is running.
don't know how you check this, but imagine it's due to python is the daemon process.
ex : download service like couchpotato, etc.
It could be an old bug, never use xbian-config for myself in fact
other is for the new xbmc gui settings,
i imagine it's under development because TODO,
but when you have a radiobutton, value must be int and not string
on 40_update.py
line 285 :
return ['1','/tmp','Daily']
should be
return [1,'/tmp','Daily']
don't ask me why i've done like this :-).
otherwise looks like you have understand the xbmc-xbian-config :-)
As i've got more or less access to a pi now,i can help a bit on this part if you need something
RE: beta2 - mk01 - 1st Oct, 2013 01:26 AM
(1st Oct, 2013 01:19 AM)belese Wrote: i imagine it's under development because TODO,
but when you have a radiobutton, value must be int and not string
on 40_update.py
line 285 :
return ['1','/tmp','Daily']
should be
return [1,'/tmp','Daily']
this is wanted. could not find other solution to make existing GUI objects greyed out (not disabled - not clickable at all).
if you put 0, is disabled. if you put 1 enabled, other number is out of range disabling whole category. string will trigger a error at log, but from outside looks as needed.
need to finish this one + simple snapshot management + copy & paste daily /weekly snapshots settings as you sent.
and with python running you mean like preprocessor for running .py scripts?
this is per service (like xbian-xbmc-bridge etc). otherwise you would have one indicator for more services. because all could be run on top of python.
xbian-config currently all possible ways of service could be started. as upstart script, as sysv script as plain daemon process or even defined via inetd wrapper.
so if you add couchpotato what is python script (if I get you right), in definition of services you can put upstart name (if there is script for it as upstart), the same name is checked for sysv (/etc/init.d) presence, if not found, process list is reviewed (if you add the daemon name). but the daemon name is not limited to what launched the script for example, you can check against process list including whole command line. so instead of "python" only you will see "python /usr/local/sbin/upstart-bridge-xbmc.py". in this case "/usr/local/sbin/upstart-bridge-xbmc.py" is defined as daemon process.
RE: beta2 - belese - 1st Oct, 2013 01:57 AM
ok,
is it just for debug/devel?
otherwise the function setEnabled exist for control.
but if you need it, i have to look deeper, didn't remember on which class it's apply
RE: beta2 - mk01 - 1st Oct, 2013 02:55 AM
(1st Oct, 2013 01:57 AM)belese Wrote: otherwise the function setEnabled exist for control.
but if you need it, i have to look deeper, didn't remember on which class it's apply
no no, don't waste time, just that object is quite big and until I get into it, add the browser for picking up block device and somehow add also the possibility to choose any folder + pick filename, will take time and didn't want to put to people looking like doing something but with no result / feedback / anything done. maybe you will help me with idea .
the text version currently expects destination as string which can be as block device directly like "/dev/mmcblk0p3" or prefixed with "file:" (e.g. "file:/mnt/file.img", in that case copy is created in img file including partition table and /boot with already adapted cmdline.txt etc.
so you can with it move your install to USB for use, or just as backup, or just over network create .img and just dd it back needed (like img downloaded from web, just it is your install 1:1).
RE: beta2 - belese - 1st Oct, 2013 03:16 AM
>the text version currently expects destination as string which can be as block device directly like "/dev/mmcblk0p3" or prefixed with "file:" (e.g. "file:/mnt/file.img", in that case copy is created in img file including partition table and /boot with already adapted cmdline.txt etc.
there are more solutions :
2 button, one for device, other for file (just herit setting class in python, and override needed function)
i can show you or even do it if you prefer this one (for example, i use it for connectivity settings in 10.system.py)
or
2 consecutive dialog, first one is a select :
either :
-yes/no dialog
but you can't change text of a yes/no button,
but you can add 3 text line to explain (yes = dev, no = file)
-a select dialog (like timezone) with choice (file/dev)
if file is select,
an extra keyboard dialog with filename
a browse dialog
if dev is select,
a select dialog with all /dev/sd* and ?)
(for me it's the more logic)
or
create a custom xml dialog,
it's possible, but more work if you want to handle button action.
the second solution is a bit like wifi_connect, in utils.py