.next alpha image


At least that confirms it really is the calibration, i have no idea why it would cause the rest of the script to fail when its run at the start.

The older script used to run it at the end, and it didn't cause any visible problems there, not does it on the zaxxon firmware.
 
This might seem a silly question, but to keep my image up to date with the .next changes, do I have to keep reflashing? or will opkg upgrade basically do that?
 
Normally the upgrade should be enough.

There was some big update about 2 Weeks ago.

You have to reflash for the big update.

That's what aTc said some tile before (or at least what I understood).

Correct me if I'm wrong.
 
Thanks both for the replies

I would recommand you to *not* flash this image just yet. install on sd card
yeah, when I say "flash", I kind of meant, "do I need to wipe my image with the new one?". I am using this off an SD card, but if I'm going to be resetting it every now and then, I'll probably not spend too much time configuring the desktop etc.
 
Thanks both for the replies

I would recommand you to *not* flash this image just yet. install on sd card
yeah, when I say "flash", I kind of meant, "do I need to wipe my image with the new one?". I am using this off an SD card, but if I'm going to be resetting it every now and then, I'll probably not spend too much time configuring the desktop etc.
I think we're not in a state where we should spend much time for desktop configuration.

Gamescom was much more work than I expected.

Didn't find a second to give the new image a try.

I hope I can test it during the next weaks.
 
Normally the upgrade should be enough.

There was some big update about 2 Weeks ago.

You have to reflash for the big update.

That's what aTc said some tile before (or at least what I understood).

Correct me if I'm wrong.
It wasn't really a "big" update, it's just that the package manager doesn't really handle files that are set up by the first-run-wizard, or config files in the users home dir.
 
Normally the upgrade should be enough.

There was some big update about 2 Weeks ago.

You have to reflash for the big update.

That's what aTc said some tile before (or at least what I understood).

Correct me if I'm wrong.
It wasn't really a "big" update, it's just that the package manager doesn't really handle files that are set up by the first-run-wizard, or config files in the users home dir.
I see.
 
I'm having problems connecting to a WiFi SSID. If I load SuperZaxxon, it connects fine, as do other devices, but it just instantly pops up saying "Disconnected" in .next; I am able to connect to other networks from .next though

How do I go about finding out what the problem is?

Cheers

Nevermind, it seems that the WPA password hadn't saved, if I sudo the connection editor and put it back in, it seems to be okay now
 
Last edited by a moderator:
I still get the unit just stopping after


[ OK ] Reached target Shutdown.
when trying to power down.
It just does nothing.

However, if I eject the SD card, it starts complaining about not being able to access some blocks, but the SD LED isn't on when it's plugged in.

So what I end up having to do is a soft reboot (pandora key + power switch) then wait for it to boot up, and then shut it down again.

This time it happened, all I did was boot up, pressed the Pandora key to show the menu, tapped down to the Multimedia to see if Audacious was there (didn't even run it), then scrolled down to Power Off->Shutdown

It's rather frustrating. Any ideas on what's wrong? Are there any logs that I could copy off to show you?
 
This only happens on mine very very rarely, and usually only when i was doing some serious messing about, so i didn't really bother to investigate deeper.

Does this happen every time you reboot ?

And does it also happen when you run "sudo systemctl poweroff" instead of going through the menu ?

Also, there are two menu options to shut down/log out, don't have my pandora here to tell you what they're called, but one of them will ask you for a password and shut down.

The other one seems to drop you back to the display manager (slim) and things sort of fail then.
 
It's never happened for me (although I've had the thing where one of the options in XFCE just dumps you back to slim).

Did this start happening after adding stuff with opkg?

If "sudo systemctl poweroff" works, you can edit what Openbox's "Power Off -> Shutdown" is calling: see line 605 of /usr/pandora/scripts/openbox-functions.sh. Currently, it's "shutdown -h now" (the sudo part is taken care of earlier).

Regarding logs, you can try running 'journalctl'. You might want to redirect its output to a text file though, 'cos it's massive. On mine, if I scroll up past all the 'login' stuff, the messages from the last shutdown are:

Code:
Reached target shutdown
Starting save random seed
Started save random seed
Started Update UTMP about system shutdown
Starting Final Step
Starting Poweroff
Shutting down
Failed to get D-Bus connection: 
Journal Stopped
 
Not every time, and I usually thought it was when I'd tried to run an incompatible PND. However, that last time I literally booted up and shut down. I wouldn't say it's all the time, but it's frequent enough to be annoying

In fact, I thought I'd try and capture it. So I booted up the Pandora. As soon as it got to the desktop, I started recording, so I only pressed four buttons (Pandora, Up, Right and Enter), and it happened on first go

https://www.youtube.com/embed/TfXI_40QSyk?feature=oembedSorry the video isn't great quality, I've only got a crappy camera

I don't get any password requests, I'm not sure what slim is, but hopefully the vid will be more helpful to you

The sudo systemctl command does exactly the same, after my softreboot I tried this, so for the 2nd time in a row, it's failed to power off. :(

Edit: Attached my journal file for you to peruse. Cheers

journal.zip
 

Attachments

  • journal.zip
    457.4 KB · Views: 442
Last edited by a moderator:
It seems pretty hit and miss, doesn't it? Sometimes it works, sometimes not. I have had it happen to me just now, bizarrely. In the absence of the other evidence, I'm blaming you ;)

Something you could try:

Enable the debug shell

sudo systemctl enable debug-shell.service
When it gets stuck at "Reached Shutdown Target", keep Alt and the Left D-Pad key to get a root shell.
Type

systemctl --full list-jobs
The one that shows a 'running' status is the one that holding everything else up (you should see things like 'shutdown' are left at a 'waiting' status until that one is complete)
For me, the 'running' process was called something like 'dev-disk-by[lots of numbers].swap', which turned out to be a symlink to dev-zram0.swap. I killed dev-zram0.swap ("systemctl kill dev-zram0.swap" and/or "systemctl --signal=9 kill dev-zram0.swap" if "systemctl list-jobs" is still showing things), and then typed "shutdown -h now".

Is it this that's stopping your shutdown too?

----

BTW - in the previous posts, don't worry about references to SLIM (which is the graphical login manager you never see if you've got automatic login enabled) and to passwords for powering off (which only happens with XFCE).
 
Last edited by a moderator:
Urm, I booted up, enabled the debug shell, and started a shutdown. It stopped in the usual place, and I hit alt-left, but all that happened was that the screen was cleared, and the cursor was positioned in the top-left spot, and doesn't respond to any input. The Pandora is still running, because the cursor flashes, but it doesn't respond to any keys.

So I soft-rebooted again, and this time I got a shell. I tried the kill commands you used, but it didn't kill the process for me. How did you find out where the symlink was? Anyway, the kill commands you gave didn't help me shutdown. If I do "shutdown -h now", it just sits there. I can however, Ctrl+C back to the root shell.

Edit:

I should have added, yes I had a "dev-disk-by" line from the list-jobs command

When I run "systemctl list-jobs", it kusts the dev-disk of type "stop" and state of "running", even after the kill :(
 
Last edited by a moderator:
It looks like the debug shell needed a reboot in order to be activated (which is unfortunate, considering what we needed it for).

For the symlink, I typed


systemctl --no-pager --full --all --type=swap
and looked at the last DESCRIPTION field (which is probably wrapped around onto the next line). It showed something like

Code:
/dev/disk/by-uuid/(long-number)
and when I typed

Code:
file /dev/disk/by-uuid/(long-number)
it showed that it was a link to dev-zram0.

In terms of killing it, I think I had to try a few times. Nothing should be able to withstand a --signal=9 kill command, but it can be a bit stubborn. I maybe did it about 3 times before "systemctl list-jobs" showed up empty. After it finally dies, 'shutdown -h now' will work.

----

Thinking about it though, if the process that was stopping your shutdown was called dev-disk-by ... .swap, then zram is the only thing it can be, 'cos we don't have any other type of swap space set up. I'll look into how zram can either be disabled from ever starting (there's a forum post about it, somewhere ...), or maybe if it can be manually stopped before shutting down, and post back.
 
'cos we don't have any other type of swap space set up.
I think I set a swap partition up on my SD card when building it, is that the same thing? Do I have to activate that partition somehow, or is the presence of it enough?
 
Back
Top