Media Director no longer boots

[m3freak]

My lone MD stopped booting a few days ago.  It still gets an IP, but when it comes to downloading the initrd and kernel, it bombs out.  Well, the actual order of fail is:

1. tries to download the kernel image that was built for it (can't)
2. then it moves on to the default (no go)
3. finally tries a local boot, but since there's no local install, the PXE boot stops.

I haven't done anything to the MD config itself.  I did however update the core.  Maybe the update did something.

I can see the tftp request on the core in /var/log/syslog.  The core answers properly (I think).  

What could be wrong?

[jwelch1324]

Did you run

Code: [Select]

sudo /usr/pluto/bin/Diskless_CreateTBZ.sh

on the core after the update? You may have to rebuild the MD images if you changed to a updated the kernel or any important packages.

[toppot]

My lone MD stopped booting a few days ago.  It still gets an IP, but when it comes to downloading the initrd and kernel, it bombs out.  Well, the actual order of fail is:

1. tries to download the kernel image that was built for it (can't)
2. then it moves on to the default (no go)
3. finally tries a local boot, but since there's no local install, the PXE boot stops.

I haven't done anything to the MD config itself.  I did however update the core.  Maybe the update did something.

I can see the tftp request on the core in /var/log/syslog.  The core answers properly (I think).  

What could be wrong?

Try looking at this thread http://forum.linuxmce.org/index.php/topic,7842.0.html

It could be either this exact error, or perhaps digging from here could lead you in the proper direction. I would assume from your description that it could be from this corrupted line in /tftpboot/pxeboot.cfg/01-<md-mac-address>

-Tony

[m3freak]

Try looking at this thread http://forum.linuxmce.org/index.php/topic,7842.0.html

It could be either this exact error, or perhaps digging from here could lead you in the proper direction. I would assume from your description that it could be from this corrupted line in /tftpboot/pxeboot.cfg/01-<md-mac-address>

That's the ticket: the pxe config file is hosed and I didn't even notice it.  Here's what mine looks like:

Code: [Select]

DEFAULT Pluto
LABEL Pluto
LOCALBOOT 0 #ERNEL 44/vmlinuz
APPEND initrd=44/initrd.img ramdisk=10240 rw root=/dev/nfs boot=nfs nfsroot=192.168.80.1:/usr/pluto/diskless/44,intr,nolock,udp,rsize=32768,wsize=32768,retrans=10,timeo=50 quiet splash video=uvesafb:mode_option=1024x768-24,mtrr=3,scroll=ywrap vmalloc=256m

If I had looked at it more carefully, I would have noticed the malformed "kernel" line.  Fail.

 I'll fix the fubar line and test out the MD boot tonight - I'm pretty sure it'll boot properly.

Do you know why this happened?  The thread you steered me to is from 2009!!!

[m3freak]

I fixed the kernel line and now the MD is booting successfully. If it happens again, I'll know where to look.

Thanks for the help!

[jamo]

I fixed the kernel line and now the MD is booting successfully. If it happens again, I'll know where to look.

Thanks for the help!

I know how it happens officially but something else might cause it as well- basically when you go to the "power" screen of your MD on the orbitor, and the MD is on, there is a button on the top right that has the legend "Restart MD as XXXXX". In my case, the XXXXX is something like "Err parameter not found".

This confused me for a while until I figured out that it is probably getting XXXXX from the MD config - see webadmin for this MD where you specify what the base operating system is for this MD. For all of mine, this field is blank, and that's probably where the "Err parameter not found" message comes from. I presume, but haven't tested, if you fill this in as "Windoze 7" or whatever, then the button will make more sense.

Anyway, it turns out that this button does fubarring of your pxeboot.cfg file and it is an intentional thing- If you choose "Restart as Windoze 7" then it writes the change you saw to the file, and restarts the machine. When the machine comes up, it tries to boot the underlying O/S which doesn't work if you don't have one!

You can test this by viewing the contents of the pxeconfig boot file before and after clicking the above button from an orbiter. You can fix manually, as you did, by editing the file or you should be able to just open another orbiter, point it to the relevant MD's power page and click this button which I guess will read "Restart as LinuxMCE" when the pxeboot cfg file is fubarred.

Does that make sense?

I discovered this over the w/end and meant to document it in the wiki because it had me a wee bit flummoxed too. When I document in the wiki I'll do it from my MCE machine so I can test and be a bit more accurate in my description... this is just from memory.

[m3freak]

Does that make sense?

Yup, that makes perfect sense.  Actually, that's a really neat feature.  I could pop a 2.5" HD in my Jetway, install Fedora + XBMC, and use that for my media needs just in the family room. When I need to use LinuxMCE, I could hit that button you described and have it boot up as a MD instead. Hmmm. 

I do like LinuxMCE and it's media integration is neat, but the presentation makes the media components hard to use.