The startup application configured to launch was missing when the RT target boot

[AMyth]

We have a deployment of 12 cRIOs. I deployed and tested the executables as startup EXEs in each of these systems. The systems were powercycled several times without a problem. We shipped these cRIOs to the customer and he turned them on after a month. 3 of the cRIOs didn't seem to be running. I connected using LabVIEW and the error I got in the connection attempt was

The startup application configured to launch was missing when the RT target booted

Of course there was the standard stuff as well.

Now I FTPed onto the targets and looked in the startup folder and it was empty. I deployed the executable, then powercycled and the applicaiton ran just fine.

The hardware we are using is the 9074 chassis with two 9205s and six 9476s.

[Logan_H.]

Amit,

I don't see much of a question here other than the implied "What happened to the 3 that don't work?" so I'll try to help solve that:

So, since you have a large batch of cRIOs (commonly known as a flock of cRIOs) we need to hone in on their differences. Is the hardware the same in each of the cRIOs? The modules installed? The version of NI-RIO?

How were the cRIOs functionality tested? Was their data checked prior to shipping or was just the LED just set to blink in the Start Up EXE?

Was the software deployed from different computers or the same? Was it deployed programatically or by hand for each of the 12 cRIOs?

Thanks,

Logan H

[AMyth]

Logan,

I was having browser/internet issues when trying to submit and it left out a good chunk of my post.

All cRIOs were deployed from the project. No replication.

All the cRIOs have the same hardware - 9074 with 2 9205s and 6 9476s.

All software will correspond to the Labview 2009 SSP1 release around february.

All cRIOs have been extensively tested before shipping. Every channel was tested. The test process lasted over a day for a pair of cRIOs.

What appears to have happen is that the cRIO remembers all its settings, just the startup EXE seems to be missing. After delivery to the customer's site, it may have been power cycled several times, but can that cause files to get deleted? Like windows does a shutdown on the cRIO lead to corruption of the file system?

There is a KB that talks about the difference between using a capital C and small c in the path for build. But if that was the case then we shouldn't have been able to set the systems as well right?

Thanks,

Amit

[Logan_H.]

Amit,

The Startup folder for a Startup RT EXE is part of the non volatile memory and therefore should stay there regardless of power loss. There are a few files that are deleted at reboot but they are part of the volatile memory.

Of course there is the Format RT Target VI that can be run by a host to format the target at a specified IP but that would leave the cRIO completely cleaned off and unconfigured.

Anything else missing or corrupt on the cRIO?

Logan H

[AMyth]

Logan,

The cRIOs were perfectly fine outside of the missing startup EXE. The IP and software was all intact. No format or any such thing was used.

Amit

[Logan_H.]

Amit,

Where are the cRIOs now? Still with the customer or with you now? Can you reinstall the Startup EXEs and do the work properly (even after multiple restarts)?

Is there any difference between the part number revision of the controllers? The revision number is the last set of numbers that are hyphenated at the end of a PN. EX: 779174-01, the 01 is the Revision #.

I've been looking through our Corrective Action Reports and found no cases of just the Startup EXE being removed from the cRIO without any sort of user interaction. Any cases of files missing have been full system corruption or more than just a single file.

Logan H

[AMyth]

Logan,

Re-deploying fixed the issue and there doesn't seem to be any corruption at all. My concern is that this will happen again...

The cRIOs are mounted and its impossible to pull them out. I can't give you much information about the revision numbers. They were all purchased together about 4 months ago.

Amit

[Logan_H.]

Amit,

Granted since the transistors in flash memory do degrade as they are written and re-written, the internal memory of the cRIO will eventually corrupt... but for a single file to go missing and nothing else to be corrupt isn't a reliable indication of this.

Is there any time in your application where your memory is dyncamically addressed and erased (even if it is the data, not the startup EXE).

Logan

[AMyth]

Its strange that this would happen just on a few systems. Its not repeated itself again. I've created a small PC based utility that deploys just the startup exe to the correct folder in case it goes missing for any reason as an auto recovery application. I can't risk the files going missing again.

Amit