[Solved] TC3.8_4 Real Time Upgrade

[MTCAT]

Hi Rich,

The drama continues, for the RT kernel, all the extensions were downloaded previously with the non-RT kernel, and then the tgz files copied to /mnt/sdb1/tce/optional.

Since then, I've actually never had the ethernet up and going with the real time kernel.

With the rt kernel, using the "base" boot option, after booting into the GUI window manager, when I type "ifconfig -a" in a terminal window, I don't even see an "eth0" at all, I only see a "dummy0" and an "l0".

Is this because we're skipping something getting loaded for eth0 with the "base" boot option ?

Thanks,

David

[Rich]

Hi MTCAT
OK, so you can load extensions. Reboot to base again. When you get to the command prompt:

Code: [Select]

sudo depmod -a
sudo udevadm triggerThen load your extensions. See if  eth0  shows up now,
Got to go. Back in about 2 hours.

[MTCAT]

Hi Rich,

You're amazing, how do you remember all these commands, that worked !, after "sudo depmod -a" and "sudo udevadm trigger" eth0 appeared out of the ether (ha-ha) and then I applied the IP address's with the Network App !

My first try at using "tce-load -i zsync" didn't work though, I received an error "zsync.tgz not found !", but, I was able to install zsync with the Control Panel App, and then for the fun of it, I tried the command line in a terminal again, "tce-load -i zsync", and then I received back "zsync is already installed !"

So, hmm, not sure why the terminal prompt command didn't work the first time, but at least the Control App did work !

Maybe try loading another extension with "tce-load" ?

Thanks,

David

[Rich]

Hi MTCAT

... My first try at using "tce-load -i zsync" didn't work though, I received an error "zsync.tgz not found !", but, I was able to install zsync with the Control Panel App, and then for the fun of it, I tried the command line in a terminal again, "tce-load -i zsync", and then I received back "zsync is already installed !" ...

Extension names end in  .tcz  not  .tgz.
Are you sure  zsync.tcz  was in your  tce/optional  directory?

... On boot-up, I wasn't given a choice between the old kernel and the real-time kernel, I just saw a little "boot: " prompt echoed to screen, and then eventually, off it went with the real-time kernel. ...

It's possible the  TIMEOUT  setting in your  extlinux.conf  file is too short for you to make a selection. Change it to:

Code: [Select]

TIMEOUT 500That will give you 50 seconds instead of only 5. There should be some way to select what to boot, Maybe the
up/down arrows?

... modprobe: can't open 'modules.dep': No such file or directory ...

Let's try to address that. Boot into the RT Kernel. When you get the command prompt:
Follow the instructions in the attached file. The new initrd will be called  tinycoreRT2.gz  so we don't overwrite the
existing one. Adjust your extlinux.conf accordingly.

Reboot and see if eth0 is found now.

[MTCAT]

Hi Rich,

Sorry I've been away from my computer all day, an annoying home improvement project, I'll boot the RT kernel, change the TIMEOUT and take a try at your new instructions to address modules.dep issue.

Sorry for my confusion re: the module extension, I just thought it was odd that the "manual" tce-load method (typed from within a terminal window) didn't initially work to install zsync, but the Control App method did work.

Also, just to make sure, with the RT kernel, I'm working from within the window manager, opening terminal windows and executing commands that way, is it fair to say that I'm technically not at a command prompt ? I wasn't sure if you thought that I wasn't getting into the window manager, and just booting to a command prompt only. I'm actually getting into the Windows manager (X ?) and opening terminal windows inside that.

Thanks,

David

[MTCAT]

Hi Rich,

I made the new initrd "tinycoreRT2.gz", and increased TIMEOUT to 500, but, not sure, was I supposed to remove the "base" modification from the APPEND line as well ?

With the "base" still in place, the RT kernel boots up into the window manager and unfortunately, I don't see eth0, and I didn't see it on the first boot up today as well (using tinycoreRT.gz), even though you helped me get eth0 going last night, and I saved the changes in the Network App manager, it still didn't show up.

I confirmed that there is a call to "eth0.sh" in /mnt/sdb1/opt/bootlocal.sh, and eth0.sh is also present in /mnt/sdb1/opt, with the correct IP address's, etc., so that seems weird.

The increased TIMEOUT worked but unfortunately, the arrow keys didn't do anything at the "boot:" prompt, so I just hit enter and off it went with the RT kernel using tinycoreRT2.gz

Thanks,

David

[MTCAT]

Hi Rich,

I removed the "base" from the APPEND line and was able to boot w/o the "modules.dep" problem, but I still see the "ramzswap0" error. Also, I saw a strange error, after loading extensions, I saw this on the screen; "ifconfig: SIOCSIFADDR: No such device and "route: SIOCADDRT: No such process"

Eventually it still booted up but with no eth0, but when I did your "sudo depmod -a" followed by "sudo udevadm trigger", eth0 was there.

Thanks,

David

[MTCAT]

Hi Rich,

Hopefully this isn't "too much information", but I attached dmesg output here as well.

Thanks,

David

[Rich]

Hi MTCAT
I don't know what the problem is. Try running the  TCscan.sh  script after the GUI is up so I can see a little more about
the state of this system. If you forgot how, you did it here previously:
http://forum.tinycorelinux.net/index.php/topic,24813.msg157696.html#msg157696
Attach the results to your next post.

[MTCAT]

Hi Rich,

Good idea, here's the ".tmp" file, had to attach it because the scan never finished, it couldn't find the bootloader config file and it just got stuck there.

Thanks,

David

[Rich]

Hi MTCAT
Since you insist on breaking my script, I've added a  "Press enter to skip."  option when it gets to the  Find bootloader part.
I've attached it here since the server won't let me update it in the original thread.

Please run it twice.
Once when you get to the command prompt.
Then:

Code: [Select]

sudo depmod -a
sudo udevadm triggerRun the  /opt/eth0.sh  script and start the GUI. Now run  TCscan  again.

Attach both files to your next post. I'll know which is which because the file names contain time stamps.
Also attach a copy of your  /opt/eth0.sh  file.

[MTCAT]

Hi Rich,

Thanks for the new script, here's the outputs, before and after getting eth0 up and running. Just to be clear, I'm doing this all within a terminal window in FLWM though, hope that's okay.

Thanks,

David

[MTCAT]

Hi Rich,

Sorry, forgot to attach eth0.sh, here it is, it looks a bit different than the non-RT kernel eth0.sh I think, I don't recall that "pkill" command for example.

Thanks,

David

[Rich]

Hi MTCAT
The only thing I see is you don't appear to have a  mydata.tgz  file. I'm wondering if maybe that is causing an error
in the  tc-config  script and messing up something else.

Let's do this again:

Code: [Select]

sudo touch /tmp/dummy
sudo echo "tmp/dummy" > /opt/.filetool.lst
Now let's run a backup:

Code: [Select]

filetool.sh -bSee if it boots up correctly now.

... Just to be clear, I'm doing this all within a terminal window in FLWM though, hope that's okay.

I don't see  flwm_topside  or anything else required for a GUI in your  onboot.lst  file.

How are you starting your GUI?

[MTCAT]

Hi Rich,

I followed your instructions below but unfortunately that didn't change the boot-up behavior, other than now loading a backup file, I still see the "raqmzswap0" error and the odd "ifconfig" error as well with no eth0 present unless I do the "sudo depmod-a" and "sudo udevadm trigger" workaround.

I'm not doing anything special to start up the Window manager, that's just what gets booted up for me somehow, this is the same thing with my non-RT version, it just always boots into FLWM.

Thanks,

David