NI-3100 no longer responsive after network failure during program upload

myer0244

Hello all,
I was updating one of my NI-3100s with new code when the network in my building went down. It is no longer responsive. I put it in PRD mode to bypass the program and still nothing. The network sees it still however, when I pull up the "Show Online Tree" in the program, it connects and shows "System 10 Devices" but nothing under it. Not even the built-in rs232 ports. I also tried going to the central controller's web page, which results in the connection timing out.

Any thoughts on what is up with the controller? Or possibly how to fix it or Factory reset it?

Any input is welcome. Thanks!

papadouk

Can you telnet the master? If yes, try to send a 'hello'. The master will respond with 'How are you doing?'.

If the above works, send a 'clean disk -f' and after you have a console prompt again send a 'reboot' and wait until your controller is up and running.

Then you should be able to transfer your program again

Kostas

myer0244

Can not telnet to it nor connect to it directly through the Program port =( However, it still maintains it's ip address. The touch panel see's it as well as Netlinx studio

papadouk

Have you tried to cycle the power? I don't think you will have an issue loosing the controller since you were not downloading firmware...

myer0244

papadouk wrote: »

Have you tried to cycle the power? I don't think you will have an issue loosing the controller since you were not downloading firmware...

Multiple times, hard power-cycle as well

papadouk

How about a firmware upload?

myer0244

Just attempted, I get this: "The master has not indicated the device is online." However, in the bottom right corner of netlinx studio, the "Netlinx-<ip address>:1319 does turn green

myer0244

This is what I see. The online device tree was just refreshed.

Bigsquatch

You didn't say if you tried to upload the program again. I assume you tried.

Maybe start a new workspace with just an empty program and try to compile & upload that.

Otherwise you may have to remove it's CF and copy over it with a working image from another controller.

DHawthorne

It's entirely possible the CF was corrupted with the network interruption. It shouldn't happen, but I wouldn't rule it out.

myer0244

Bigsquatch wrote: »

You didn't say if you tried to upload the program again. I assume you tried.

Maybe start a new workspace with just an empty program and try to compile & upload that.

Tried both of these, and both uploads failed. The transfer process wouldn't even start.

Bigsquatch wrote: »

Otherwise you may have to remove it's CF and copy over it with a working image from another controller.

Don't know how to do this.

PhreaK

myer0244 wrote: »

Don't know how to do this.

If you open up the master you'll see a compact flash card in a card slot. You can remove this, put it in a card reader on your computer a do a direct copy (not just copy/paste of files - but dump an image onto it) of a functional card. There's step by step instructions on how to do this floating around the forums (sorry don't have the link handy).

myer0244

***UPDATE*** (still not resolved though)
I have deduced that the problem does not lie within the central controller, but actually in the network. The "IPDeviceDetector.run" normally detects multiple other devices. But after the big network problem on Friday, it no longer does. The above image is what it looks like when it is trying to boot. The image attached to this post is what it is supposed to look like.

I came to this conclusion first by replacing the central controller with a "Known working" unit, which booted up with the same result, but worked 10 min previous at a different location. I figured since it was an issue wit h the network connection during the boot sequence, I unplugged the network cable, fully booted the central controller, then re-attached the network cable, and it worked perfect. However when I reboot it when it is still attached to the network, it hangs at that one point.

My question is: what in the boot sequence (that is network-related) can hold the program from running? and is there a way to disable it. (such as the multicast IPDeviceDetector)

-Justin

jjames

As far as my tests have gone, you can completely kill Duet and still have it boot up fine. (just no Java related stuff will work.) When I get home I can run some tests. There's very little (if anything) that you can modify that's part of VxWorks startup.


Sent from my iPhone using Tapatalk

kishormohiite

Hi guys,

I am struggling to transfer files to NI-3100 for last few hours. I tried it 6-7 times. But it fails unexpectedly. Though, I can see the panel online with all ports and devices intact in online tree. Also gets responses from connected devices. And all panels connected to it are working perfect. Only issue is with uploading new programs..
Can anyone help me with solution to fix this please.

kishormohiite

Addition to above:
if I use > @papadouk said:

Can you telnet the master? If yes, try to send a 'hello'. The master will respond with 'How are you doing?'.

If the above works, send a 'clean disk -f' and after you have a console prompt again send a 'reboot' and wait until your controller is up and running.

Then you should be able to transfer your program again

Kostas

Hi,

if I use this 'clean disk-f' command, will it erase all programs on NI-3100? .

HARMAN_icraigie

@kishormohiite said:
Addition to above:
if I use > @papadouk said:

Can you telnet the master? If yes, try to send a 'hello'. The master will respond with 'How are you doing?'.

If the above works, send a 'clean disk -f' and after you have a console prompt again send a 'reboot' and wait until your controller is up and running.

Then you should be able to transfer your program again

Kostas

Hi,

if I use this 'clean disk-f' command, will it erase all programs on NI-3100? .

Yes, when the system reboots there will be no code to run.