Snakeoil Forums

Full Version: read_cfg problem when running 64 bit 0.1.4 on atom system
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Pages: 1 2
Hi AK, thanks for your work on snakeoil.

I can run the 32 bit version of snakeoil 0.1.4 just fine, but the 64-bit version is giving me problems, including:

500 - Internal Server Error
[Image: sad.png]

It seems that the audio side of things is running just fine, but whenever I change any setting via your web interface, your read_cfg executable is crashing and then apport processes chew up the CPU for a long time before everything eventually comes back to normal.

/var/log/apport.log is full of the following:
 ERROR: apport (pid 26297) Mon Jun  6 20:45:21 2016: executable: /var/www/cgi/read_cfg (command line "cgi/read_cfg 12345")ERROR: apport (pid 26297) Mon Jun  6 20:45:21 2016: is_closing_session(): no DBUS_SESSION_BUS_ADDRESS in environment

As I mentioned, I only see this with the 64-bit version; 32-bit runs fine!

Please advise.
P.S.: For completeness, here is the tail of dmesg output:
 nit: systemd-logind pre-start process (1927) terminated with status 32init: avahi-cups-reload main process (2105) terminated with status 1init: Failed to obtain startpar-bridge instance: Unknown parameter: INSTANCEe1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Txinit: smbd main process (2074) killed by HUP signalinit: smbd main process ended, respawninginit: samba-ad-dc main process (2186) terminated with status 1init: failsafe main process (2019) killed by TERM signalinit: network-manager main process (2312) terminated with status 127init: network-manager main process ended, respawninginit: plymouth-upstart-bridge main process ended, respawningclear_cache (14812): drop_caches: 3
(06-Jun-2016, 08:51 PM)rab Wrote: [ -> ]500 - Internal Server Error
[Image: sad.png]

It seems that the audio side of things is running just fine, but whenever I change any setting via your web interface, your read_cfg executable is crashing and then apport processes chew up the CPU for a long time before everything eventually comes back to normal.
So any setting change would cause that problem?

Can you check how much free space you have on /? You can find that information in the Music tab.
1. when setting player or music mount (prob because that resets player...?)

2. plenty of space:

[Image: Clipboard_image_2016_06_08_08_29_57.png]
If you look at the first row ("Free"), you'd see you are booting from your USB (or CD media). And that tells you it is 100% full (another way of saying it's a read only filesystem).

If you have already installed Snakeoil on the 120 GB SSD (I think you did), remove the boot USB stick from your computer, and then turn on the computer. When you look at the music tab again it should show /dev/sda1 mounted as /. and your free space should be < 100%. And then everything should work.
Indeed, I am booting from USB, as I don't want to install it on my SSD until I confirm that it works.
If you look down to the third entry, you'll see there is 3.9GB free on /, so plenty of space there.

In any case, the same applies when I boot the 32-bit version, which runs just fine from the same USB key.

Thoughts?

- r
(08-Jun-2016, 05:06 PM)rab Wrote: [ -> ]Indeed, I am booting from USB, as I don't want to install it on my SSD until I confirm that it works.
If you look down to the third entry, you'll see there is 3.9GB free on /, so plenty of space there.

In any case, the same applies when I boot the 32-bit version, which runs just fine from the same USB key.
It wouldn't be able to write to that filesystem even when there is space available, as it is a read only filesystem.
(08-Jun-2016, 05:06 PM)rab Wrote: [ -> ]In any case, the same applies when I boot the 32-bit version, which runs just fine from the same USB key.
That's interesting, can you take a screen shot of the music tab screen so I can see what partitions are mounted for 32 bits?

Edit: Actually, since it is a liveCD, you should still be able to write, but the files will be lost on reboot.. Can you show me the Home screen as well? I'm curious to see the memory stats you have when in 64 bit mode
OK, here's the 32-bit music screen:


[Image: Clipboard_image_2016_06_08_21_51_05.png]
Hmm this is a strange one.. I can't recreate this issue you have on my machines.

Is there some specific steps you do that will cause the crash, or really clicking Save anywhere will cause this behaviour?

Here are some things you can try: 
  • Redownload the ISO
  • Try the USB on anoher machine and see if you get the same issues
  • (long shot) Try another USB drive 
  • (long shot) Connect a monitor to the machine, and hit CTRL-F1 to get to the console. Now make any change to the settings and see if any errors turn up?
What software did you use to write the ISO to your USB?
... and here's the 'home screen':


[Image: Clipboard_image_2016_06_08_22_42_37.png]

(08-Jun-2016, 08:06 PM)agent_kith Wrote: [ -> ]Is there some specific steps you do that will cause the crash, or really clicking Save anywhere will cause this behaviour?

Yep, clicking save after selecting the USB interface triggers the error.
 
(08-Jun-2016, 08:06 PM)agent_kith Wrote: [ -> ]Connect a monitor to the machine, and hit CTRL-F1 to get to the console. Now make any change to the settings and see if any errors turn up?

You mean control-Alt-F1... No errors there, but i already posted the triggering error in my first post...

***

OK, I had assumed that you had written is_closing_session(), but now I realise that is not the case, and indeed I am not the first to see these error messages; see, for example:

http://ubuntuforums.org/showthread.php?t=2194358

which suggests it is a Light Display Manager issue.
(08-Jun-2016, 08:43 PM)rab Wrote: [ -> ]No errors there, but i already posted the triggering error in my first post...
appport is a crash reporter. I'm more interested in the system logs... it will show up in the console, or you can just run dmesg from the command line.

Looking at the console is probably easier as you can correlate any output messages at the time when you click the save button.

Another source you can look at is /var/log/lighttpd/error.log. Cut and paste the last few lines so I can see what's happening.

(08-Jun-2016, 08:43 PM)rab Wrote: [ -> ]which suggests it is a Light Display Manager issue.
LightDM is the graphical manager. When you use the web interface, you're iteracting with the web server. In general the two can't interfere with each other unless there is an underlying problem (and it can be anything from a bad download to hardware issues like bad RAM).

Have you tried the 64 bit USB on another computer and see if it works?
Pages: 1 2