Strange Media Director PXEBOOT behavior - FUbar'ed

[finsdown]

Ok, got a strange one here, where I PXEBOOT a pc and everything goes Haywire. I'm at a loss to why. 

Here's the situation:  I deleted one of my MDs because I kept on getting a video error message with the Nvidia drivers.  I went into webadmin and deleted the MD from the system.  Then I PXEbooted and the pc did the usual thing about announcing itself to the Core. Next it runs the DHCP_?.sh script, then the NFS? script.  Immediately after that it reboots and repeats these steps over again and again and again. 

Meanwhile on the Core/Hybrid I get a message saying "this seems to be a new orbiter, shall I set it up for you?"  Next the LMCE babe video starts up. Then I go into the LMCE webadmin tool and notice there is a new orbiter listed there, and also a new Media Director device created. The room listed for the new MD is same room my Core is in.  Well I already have a MD at that location. So it appears that LMCE is trying to create another one there.  This seems to confuse everything including me.  I tried changing the room location for it to where I wanted it to be but it won't stick.  I also get messages telling me such and such device #183 is invalid, shall I delete it. Hell yea, delete it for me.  Then once I delete the rabid Media director and the wayward orbiter in webadmin, things seeem to settle back down.   

So I'm wondering what has happened.  Did I screw something up when I deleted the MD?  I'm now unable to re-establish the MD I wanted to put back in the system. 

Thanks,

fins

[darrenmason]

Maybe re-run the "Setup Diskless Media Directors" button?

Maybe first, try deleting the old MD's filesystem from your core /home/diskless/XX or something like that.

HTH
Darren

[finsdown]

I think with 0710 that using the "Setup Diskless Media Director"  is no longer used??? Not sure. I tried that but it didn't work. Also the /home/diskless/XX filesystem for that device # is not there, so I'm assuming it's gone. I think the key to my problem is with the Media Director and it's continuous rebooting.  It starts the PXE boot fine, announces itself to the Core, starts running the installation shell scripts, and then reboots, and starts this whole process over again. I noticed that it doesn't even create the /home/diskless/XX filesystem at all, even though a device # is assigned to it when the Core detects the new MD. However, in LMCE admin the new MD is listed.  Some might say just rebuild the Core, but I've got too much time invested in getting the Core running to do this (only as a last resort anyway).   

Thanks,
Fins

[colinjones]

finsdown - long shot, but when you PXE boot to start building your MD (after you have deleted the old one) what is the first IP address it gets from the core? Is it getting one from your "unknown" devices range (I think this is the higher range) or the lower "MD" range of IP addresses?

If it is the lower one, this could be the problem - the Core maybe remembering the MD from its MAC address and putting it straight into the MD range which in turn may be screwing up the auto discovery scripts in some way I don't understand. Normally, it should get a high IP address as it doesn't recognise the device, builds itself, then reboots and gets a lower address.

[finsdown]

colinjones - yes it is getting the lower ip address 1(92.168.80.2) and I think your are right on about the core remembering the mac address from the previous MD.  Now how I wonder if  I can tell the core to not do this?  I thought maybe I could use a different network card, since I do have some spares, and that would be get around it. But I'd rather fix the problem than work around it.  Ideas?

[colinjones]

It may not be causing the problem... but if you look in /etc/dhcp3 you should find the config file and that contains the leases against MACs. You can delete it from there. Now in my experience this doesn't always stop it from remembering.... don't know why cos it is definitely related to it. But sometimes I have had to dick around a bit before it "forgot" but that was with PC shares where the DHCP client actively requests the old address. With PXE I doubt very much if this is stored in CMOS so maybe it will be easier for you...

[Zaerc]

You might have to do some cleaning up under /tftpboot/ as well.

[finsdown]

Thanks for the ideas. I will try these and post my results.

[royw]

Under Wizards, Media Directors, I did a "Rebuild Image" on my diskless MD when it was constantly rebooting.  That solved the problem for me.  I did have to go thru the AV setup wizard again.

HTH,
Roy

[jondecker76]

Wow, this is the EXACT thing that is happening to me - only I have never tried to install another media director before (I only have a core/hybrid right now, and never attempted to install any other md's). THis was the first media director I was going to PXE boot - and on the first attempt onward, I get the same behavior - Announces itsself to the router, then I can quickly see a DHCP...sh and NFS script, the word Rebooting.. than pow, it reboots and the process starts over. Mine was also assigned an IP of .80.2

[finsdown]

Good to know I'm not alone in the wilderness.  I haven't had time to try these ideas yet. I've been busy with other issues on my other MD's. I have a Core/Hybrid in the LR and a MD in the MBR that are working fairly well, except the frequent lockups with MythTV.

I will post my results when I get around to this MD again. 

Cheers -

Fins

[jondecker76]

Here are the last few things I see before the system reboots. (and the process starts all over again)
(I had to take a picture of the screen with a digital camera and type it all out...)

Code: [Select]

....
....
....
[  113.151024] e1000: eth0: e1000_watchdog: NIC Link is Up 1000Mbps Full Duplex, Flow Control: RX/TX
IP-Config: Eth0 guessed broadcast address 192.168.80.255
IP-Config: Eth0 complete (from 192.168.80.1):
  address: 192.168.80.2        broadcast: 192.168.80.255      netmask: 255.255.255.0
  gateway: 192.168.80.1        dns0       : 192.168.80.1         dns1     : 0.0.0.0
  rootserver: 192.168.80.1 rootpath:
  filename  : /tftpboot/pxelinux.0
Begin:  Running /scripts/nfs-premount ...
IP:  192.168.80.2; MAC:  00:1B:21:10:99:A5  ; Gateway:  192.168.80.1;  Architecture: i386;  DT:
***********************************
* We announced ourselves to the router. *
***********************************
Assigned permanent IP '192.168.80.2' to device '60'
Running Diskless_Setup.sh
[  116.840862] scsi 0:0:0:0:  Direct-Access        HP     Photosmart C3180 1.00 PQ: 0 ANSI: 2
Running DHCP_config.sh
Running Diskless_ExportsNFS.sh
Rebooting
[  124.271625] md: stopping all md devices.

the part that is really confusing is this:

[  116.840862] scsi 0:0:0:0:  Direct-Access        HP     Photosmart C3180 1.00 PQ: 0 ANSI: 2

The HP Photosmart C3180 is my printer... What does that have to do with PXE booting??

[hari]

(I had to take a picture of the screen with a digital camera and type it all out...)

gives a new sense to the word hardcopy

did it create a file like that:

Code: [Select]

hari@dcerouter:/tftpboot/pxelinux.cfg$ more 01-00-19-de-ad-be-ef
DEFAULT Pluto
LABEL Pluto
KERNEL 68/vmlinuz
APPEND initrd=68/initrd.img ramdisk=10240 rw root=/dev/nfs boot=nfs nfsroot=192.168.80.1:/usr/pluto/diskless/68

?

best regards,
Hari

[Zaerc]

Can you see the MD's files under /usr/pluto/diskless/# and/or /tfptboot on the core?  Does it have an entry in the web-admin?  If so you can try letting it rebuild or if not you could try adding it manually and specify the MAC and IP addresses.

[jondecker76]

Here are some results after trying some suggestions:

Code: [Select]

linuxmce@dcerouter:~$ :/tftpboot/pxelinux.cfg$ more 01-00-19-de-ad-be-ef
-bash: :/tftpboot/pxelinux.cfg$: No such file or directory
linuxmce@dcerouter:~$ :/tftpboot/pxelinux.cfg  more 01-00-19-de-ad-be-ef
-bash: :/tftpboot/pxelinux.cfg: No such file or directory
linuxmce@dcerouter:~$


Code: [Select]

llinuxmce@dcerouter:~$ cd /usr/pluto/diskless
-bash: cd: /usr/pluto/diskless: No such file or directory



Code: [Select]

linuxmce@dcerouter:/$ cd /tftpboot
linuxmce@dcerouter:/tftpboot$ dir
cnu70.2-7-4-134.sbn    Jar70.2-9-0-117.sbn    term70.default.loads
CVM70.2-0-0-112.sbn    pxelinux.0             United_States
default                pxelinux.cfg
English_United_States  TERM70.7-0-1-0s.loads


Also, there are no MD's listed on the web admin (other than my Core/Hybrid)

Something just hit me.....
The computer I am trying to network boot is the computer I normally use for my personal computing. It is running Ubuntu 7.10. I installed the XP version of the orbiter on it, and I run it through wine. The device ID listed on the orbiter page is 60. Is this what is causing the problem?
What about the HP Photosmart printer appearing in the PXE boot text - what is that all about?