Stack Full error. Control Engine stack has grown to big

Anyone else seen this error before? I have a started a ticket with support already.

When I hit abort, the PAC controller has it’s memory totally cleared. Strategy & Persistent data gone. When I hit ignore, it keeps on moving, strategy keeps running and persistent data is fine. Next time this happens I will do ignore and then restart the PR1 to see if everything is still there.

I will update this when I have some answers from support.

Yes, I have seen it as well.
We also pushed it to support, so hopefully they are working on a fix or at least getting to the bottom of what it is.

1 Like

Thanks Ben,

Did you notice any problems besides this showing in PAC Control. For example, if you just let it run did things seem to keep working? I’m on a time sensitive project but I can’t send this out if it is going to ‘die’ on us.

Unfortunately I was in a pretty fluid situation with it and could not look at the logs on the PR1.
I saw the error in PAC Control like you, so I was not sure what the long term effects of it will be if you are not in debug mode.

@darbeau Were you also running a groov View project when you got the error?

@Beno I was running a very basic project. But it did have my ‘local’ controller as a device. I was using node-red to poll a modbus slave and insert the data into a pac table. I was polling 11 devices and inserting data into 20 tables based on floats or integers. I was also subscribed to a MQTT server and inserting ‘relevant’ data.

At first I thought I was just sending to much data every 5 seconds, but I put an RBE block on it, and slowed down the writes considerably, but still the problem happened randomly.

The extra info might be extraneous, or it might help form a bigger picture.

It helps. I came across this independently here, and it looks like an issue with how groov View is reading numeric tables. Do you have your groov View project displaying the values in those tables?

Yes I do read from numeric tables. How about writing to numeric tables?

My temporary workaround would be to create a bunch of float/integer variables and read/write them that way. Have PAC ‘translate’ those to/from tables so I don’t have totally change the way things work.

I use tables to cut down on the amount of reads/writes in Node-Red

Engineering found the cause of the issue is related to using Groov View to read numeric table values from either a GRV-EPIC-PR1 or PAC controller. This affects Groov View R4.1a and R4.1b and PR1 firmware 1.3.0

If not in Pac Control debug, are there any issues to allowing this to keep running? Or am I going to need to downgrade firmware to meet my project deadlines.

Writing should be fine.

I don’t think so, but Ben might know better. For me, I didn’t even have PAC Control running, just talking to an R1, and it recovers pretty much instantly. I see a blip of yellow triangles in groov View at regular intervals, but everything keeps working.

I do see the blip every once in a while, that I can live with. But when in Pac Control, if you hit abort, it clears the memory including persistent data.

Thanks for the replies.

I just had this happen on my S1. Very disturbing. As noted, all of my data, including persistent variables were erased.

It happened when I was loading an updated strategy.

If you’re running into this, please contact Product Support. There are betas available for groov Server for Windows, GROOV-AT1, GROOV-AR1, and GRV-EPIC-PR1 that address it.

This happened on an S1 while updating the project, is it the groov overrunning the S1/causing the S1 to crash?

I just got access to the beta firmware. Will be trying it out shortly.

If your groov is reading numeric tables on the S1, yes.

FTP server is sloooooowwwww! I am averaging 30-40kb/s

There’s something going on on your end. I’m averaging over 200 kb/s.

So far everything is working fine. I haven’t gotten any more messages in PAC Control. I will keep it going and let you know if there is any changes to this.

I found an issue and what I had to do to resolve it. Not sure if this is related or not.

When reading from a numeric table and you read from an index greater than what is in the table, it stops communicating with the PAC controller. For example my table tank_levels had a table length of 10, but I was reading the 11th index. Since it was not there the problem happened.

In PAC the error table shows -12, invalid table index. Once I made the table bigger and uploaded and refreshed the groov page, it was all working again.

I understand why this would throw an error, but I think it could be handled more gracefully with warnings about the table being to small or something like, not just that it can’t communicate with the controller.

I have attached a screen shot of groov.

Groov is not crashing or giving the stack full error,

but this is happening instead. There are no error messages.

Just ‘no communication’

It has happened to me about 3 times in the last 1 hour. On pressing abort and then clicking run to start strategy, there is no loss of persistent

data. But the PAC control engine does stop.