Worflow Execution had an error: Cannot read properties of undefined (reading 'nodeExecutionStack')

[puneet-10xers]

Hey Team,

We are running a workflow for Data fetching.

When we run this workflow unexpectedly it stops after some time(ex: 120 mins, 94mins, 238 mins) without giving any details on why it stopped with an error

Below is the screenshot of the Error msg it displays.

We have deployed n8n instance to AWS EC2 on t3.large which has 8GB of RAM.

Even I have tried with setting up the Environment Variables so that it prunes the data and doesn’t impact much on memory and CPU.

sudo docker run -it -e N8N_SECURE_COOKIE=false -e EXECUTIONS_DATA_PRUNE=true -e EXECUTIONS_DATA_MAX_AGE=24 -e EXECUTIONS_DATA_PRUNE_MAX_COUNT=1000 -e EXECUTIONS_MODE=regular -e EXECUTIONS_DATA_SAVE_ON_ERROR=all -e EXECUTIONS_DATA_SAVE_ON_SUCCESS=none -e EXECUTIONS_TIMEOUT_MAX=300 --rm --name n8n -p 5678:5678 -v n8n_data:/home/node/.n8n docker.n8n.io/n8nio/n8n >output.log 2>&1 &

Can you guys help me to understand this problem and rectify it?

Thanks in advance.

Best regards,
Puneet J

[krynble]

Hello @puneet-10xers

Sorry to hear you’re having trouble.

I see the docker command you used is forwarding the logs to output.log - do you have any more data in there?

Also is it possible to add another environment variable N8N_LOG_LEVEL=debug so that n8n outputs more data? This could help us identify what happened in this specific scenario. The screenshot alone at this point just leads me to believe n8n had an issue saving the execution, but it’s hard to say why this happened at this point.

[puneet-10xers]

Hey @krynble ,

Thanks for helping out. Sorry for the delay in response.

Yeah, I am attaching a few screenshots from n8n.log file which might help us in the scenario.

Thanks in advance.

Regards,
Puneet J

[puneet-10xers]

Hey @krynble ,

If you want the whole log file, you can find it in the drive location here -
n8n.log

Regards,
Puneet J

[krynble]

Thanks for the additional information @puneet-10xers

By the logs, I see that the first image shows that n8n was trying to resume the execution with ID 4 in the very beginning. Then it failed to do so because nodeExecuteionStack was not readable.

Il’l have to ask @Jon to direct this to the rest of the team, n8n is probably having issues saving the execution when it’s put to wait and then when resuming, it fails.

Maybe with this information the rest of the team can address it.

[puneet-10xers]

Sure @krynble. Thanks for the quick response.

I will wait for @Jon to provide further guidance and additional information.

[puneet-10xers]

By the way @krynble does this issue never happened in a non-self hosted version? If yes then how do you solve it?