Forum

how you will trigger manual action start when we don't have buttons there ? or you call button the on/off switch?


[for planned / reccurent]
"the auto image" on/off is ok there
path/filename below ok
/ or /
let's change block device to LABEL (as block device will change over time, label stays - I will lookup block device based on that . at the same time LABEL will be reused to is ok for recurrent event
daily / weekly ok below

this block is fine. I would welcome if you could finish the UI workflow there + above it option for manual / one-shot copy - it is up to you, you managed the whole story before you will do good again.

I'm going to do snapshots. will do new category. in it three lines like is "delete service" for instance. the 3 options would be destroy / rollback / create ad hock. above it will copy & paste the regular planned as delivered by you.
rollback / destroy will open like TZ dialog with listing and create will open list of sub volumes to choose from. best would be like multiple checkboxes, but my moto for today is keeping it simple.

Quote:and with python running you mean like preprocessor for running .py scripts?

i was thinking it, but it's looks like it's more general, some service (even non python) have stopped status but was running.
i was first thinking about python because all python service have status off while running.


ex : couchpotato
ps -alx
http://pastebin.com/AV4ASage

xbian-config services status
sudo xbian-config services status
avahi-daemon 5
console-setup 2
couchpotato 3
cron 5
dbus 5
gssd 3
headphones 3

value :
5: Running and autostart enabled
- 4: Running but autostart disabled
- 3: Not running but autostart enabled
- 2: Not running and autostart disabled

---

>how you will trigger manual action start when we don't have buttons there ? or you call button the on/off switch?

i'm no sure i follow you. for a manual action,yes, you need a button
a button like 'manual backup'

EDIT : ok understand
yes it could be done like this, or for multisetting control, on setxbianvalue is called when all group lost focus,
we can do a check there is all info is valid, if no, ask user to enter a valid path/name,...
something like i do for static ip.
iif i remember well, there is already a function to override to check user validity data.

to resume :

gui is like this???

-----------------------
radio button : enable auto_image
path (or file) : destination
text? : backup name (visible if destination is not a device)

button : do a backup now
-----------------------------
check about validity (and save) is done when first group lost focus.

just 2 rem :
whith this, it looks possible to auto-backup to a device. is it make sense? or if it's a device that is selected, it automatic a installation (and future boot?) on this device.

I'm just upgrading to the latest release - and before the ssh "apt-get upgrade" has completed, I get a "Do you want to reboot" pop-up in xbmc.

Surely this is not correct, as if a user says "yes" then the upgrade will be halted - with a mish-mash of versions and more than likely not reboot correctly if at all?

>I get a "Do you want to reboot" pop-up in xbmc.

yes i've got too, but it looks correct, i say yes and it reboot and all went fine and updated.

(1st Oct, 2013 05:04 AM)belese Wrote: [ -> ]>I get a "Do you want to reboot" pop-up in xbmc.

yes i've got too, but it looks correct, i say yes and it reboot and all went fine and updated.

Ok - Fair dinkums. I'm not convinced! As a lot more happened in the upgrade after that message.

For me, can you do another apt-get but with the -s option, apt-get -s install and see if anything needs to be updates. TIA.

0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded.
i don't really pay attention where it was during installation.

(1st Oct, 2013 05:36 AM)belese Wrote: [ -> ]0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded.
i don't really pay attention where it was during installation.

Damn that "Slovakian dark wizard" he IS the dogs cojones.

I must admit the upgrade is looking sweet.....

(1st Oct, 2013 04:57 AM)IriDium Wrote: [ -> ]I'm just upgrading to the latest release - and before the ssh "apt-get upgrade" has completed, I get a "Do you want to reboot" pop-up in xbmc.

Surely this is not correct, as if a user says "yes" then the upgrade will be halted - with a mish-mash of versions and more than likely not reboot correctly if at all?

really funny. how for weeks I'm asking you for apt-get update from cmd line (because of all the small bugs and glitches caused each run with xbian-config to miss something. finally, when nothing has been left behind, when all cases statuses are followed checked, when at the end you can stick to XBMC config or xbian-config CLI … you starts using apt-get update ?

anyhow. it is like this: there are packages and packages. all could be updated live without special care (like stop xbmc or something else). there are some which will not be in effect until reboot. those during it's install trigger reboot-required. this is ignored silently until apt-get or dpkg finishes. then it depends. if you are in xbian-config GUI or XBMC is started, you are notified. If you decline, you won't be bothered again. If nor XBMC nor xbian-config GUI is running, reboot is scheduled automatically in 1 minute. can be canceled by "shutdown -c". those few packages are xbian-config-shell (because it's influence over initramfs.gz, initramfs.gz itself, of course kernel and finally XBMC. all other you will barely notify.

---

(1st Oct, 2013 04:57 AM)IriDium Wrote: [ -> ]I'm just upgrading to the latest release - and before the ssh "apt-get upgrade" has completed, I get a "Do you want to reboot" pop-up in xbmc.

Surely this is not correct, as if a user says "yes" then the upgrade will be halted - with a mish-mash of versions and more than likely not reboot correctly if at all?

of course we would be stupid to allow that but firstly the need to reboot is checked. if true, then running apt-get & dpkg is checked.

at least that was how I designed it. there is although a second place where reboot need is checked - during closing of xbian-config-xbmc. will add that second check also there.

btw: apt-get -s upgrade (for checking) will do for manual control, because you see the KEPT BACK packages on the screen. but in general better use


the other benefit is it runs fast as hell. 2s max.

(30th Sep, 2013 03:23 PM)mk01 Wrote: [ -> ]

Code:

8) removed /home and /lib/modules from /etc/fstab

I didn't have a separate /lib/modules entry there. Also - could you tell me what to do with /etc/fstab.ext4 and what it is used for / why it is separate from the usual /etc/fstab?

(30th Sep, 2013 03:23 PM)mk01 Wrote: [ -> ] forgot important info: with kernel is configuring IP and doing interface up on it's own during boot, /etc/network/interfaces must not contain "auto eth0" and "iface inet" must be set to "manual" instead of "dhcp/static". if interface is not auto system will not be waiting for the particular adapter to appear and will not try to deconfigure/reconfigure it during "ifup -a" "ifdown -a" - it will boot, but probably you end up with slow boot (root being disconnected / reconnected) or failed boot (missing network card although needed to be auto up) etc.

I didn't fully catch that - is that something I still should configure?

(1st Oct, 2013 04:34 AM)belese Wrote: [ -> ]ex : couchpotato
ps -alx
http://pastebin.com/AV4ASage

belese, yes, I got the point. those are the same as xbian-bridge-xbmc. to work within new possibilities / workflow, this is record in services for it:

so it is called "xbian-xbmc-bridge", it checks for existence of /etc/init/xbian-xbmc-bridge.conf or /etc/init.d/xbian-xbmc-bridge and running status from status of it's upstart or sysv script (if found) otherwise from process list by looking for "/usr/local/sbin/upstart-xbmc-bridge.py" -> where "python /usr/local/sbin/upstart-xbmc-bridge.py" will match.

you put the /usr/local/include/xbian-update/config/services lines for those services manually in or how? the install packages are adding them?

then it needs to be updated.

btw: I would ask you for last favor. in the 40_updates python, you keep available updates in array and removes always the one, which was installed. but, if updating xbian-update which has dependencies, apt-get with install it during one round so if xbian-update is to be updates and xbian-kernel and xbian-update depends (or breaks), as xbian-udpate is installed, kernel is as well. but it will not be removed from the list because we don't know that it was installed as well.

because xbian-apt-show-versions -u runs max 1-2s can you simply after ANYTHING is installed just clear the two arrays and run xbian-config update list packages | list upgrades ?

(xbian-config) removes the /tmo/aptstatus always after apt-get install round so if you run the lists xbian-config runs xbian-apt-show-versions -u and you get fresh lists.

>you put the /usr/local/include/xbian-update/config/services lines for those services manually in or how? the install packages are adding them?

no idea, package i think, curlymo should now... but i've add nothing manually.

for update will look, should be possible, as i already reload service when new package is installed but don't remember how i was doing that, will take a look

and for backup, i've my idea for gui.
should have a bit time this week to look a this.

(1st Oct, 2013 07:04 AM)wind-rider Wrote: [ -> ]I didn't have a separate /lib/modules entry there. Also - could you tell me what to do with /etc/fstab.ext4 and what it is used for / why it is separate from the usual /etc/fstab?

I didn't fully catch that - is that something I still should configure?

fstab.ext4 is just backup from auto conversion to btrfs.
beta2 has independent /lib/modules to root, so if you are on beta1.1 you won't have it.

to the second point, I updated the original post, please read again.

But if you are on Beta1.1 you won't get it running. You haven't told that? I was somehow assuming (being in BETA2 thread) it's clear you have it.

The reason for not possible is missing kernel support for NFS (there are modules). So you need go with initramfs, but this was slightly broken and missing mount.nfs. I corrected this, but again on BETA2 devel stream so you won't get it back to your install.

But I see no reason not to run the upgrade.

(30th Sep, 2013 03:23 PM)mk01 Wrote: [ -> ]

Code:

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/

I tried this after performing the previous steps, but when I did this as user "xbian" I couldn't move all files because rsync complained a lot about 'permission denied (13)'

when I performed 'sudo rsync -avx --progress --delete / /mnt/nfs_root' I got a lot of errors like 'rsync: chown "/mnt/nfs-root/etc/fonts/" failed: Invalid argument (22)'

(1st Oct, 2013 07:24 AM)belese Wrote: [ -> ]>you put the /usr/local/include/xbian-update/config/services lines for those services manually in or how? the install packages are adding them?

no idea, package i think, curlymo should now... but i've add nothing manually.

for update will look, should be possible, as i already reload service when new package is installed but don't remember how i was doing that, will take a look

and for backup, i've my idea for gui.
should have a bit time this week to look a this.

btw: I committed all my updates . so it's on git.
you have this part of code 40_update.py line 125. you simply removes this one object from "updates" class on success rc from xbian-config updates progress. then "notify success" follows so I suppose instead of removing one, you can reInit the class (or remove all updates) and after NorifySuccess you can run xbian-config list upgrades and list packages. instead. or if you do it before notify success, until user reads the box, list (if any packages left) is updated again. But better do it you.

Add new control & objects is "easy story" cause not affecting others. But this is quite long continuous workflow.

(1st Oct, 2013 07:25 AM)mk01 Wrote: [ -> ]But if you are on Beta1.1 you won't get it running. You haven't told that? I was somehow assuming (being in BETA2 thread) it's clear you have it.

I'm sorry for not making this clear - I asked it here because I knew here new functionality was being tested / developed. I'm upgrading to beta 2 now.