Unexplainable restart of container in ACI

Tim van der Heijden 1

We create PDFs in de Docker container on ACI (and other stuff to support said process). Everything works as expected when 300-400 PDFs need to be created. However, when we increase the number the following happens: in the begin of the run we can check the logs and everything is fine, after a while when we click refresh for the logs, the container restarts. The memory usage is stable around 300-500mb and CPU usage is stable. The metrics don't imply issues with the resource, but maybe there are. What can cause this issue and how can we check it?

Properties

Image on ACR
Ports: 80
Protocol: TCP
CPU cores: 2
Memory: 5 GB
GPU SKU: None
GPU count: 0
OS: Linux
Restart policy: Never (we trigger it with a timer)
Location: westeurope
IP adress type: Private
Using a network profile with private IP, NAT gateway, VPN + subnet

srbhatta-MSFT 8,546 Reputation points Microsoft Employee

2022-10-28T12:21:39.747+00:00

Hello @Tim van der Heijden ,
Welcome to Microsoft QnA.
Could you paste the entire container logs?
By how much are you increasing the number?
Tim van der Heijden 1 Reputation point

2022-10-28T15:11:38.093+00:00

Hi @srbhatta-MSFT , thanks for your response.

We approximately double the number.

We do not have the logs of the container that restarts because those are lost right? Or do we miss something in our setup?
Felipe GARCIA 0 Reputation points

2024-04-11T12:12:45.02+00:00

Hello @Tim van der Heijden @srbhatta-MSFT ,I know I am arriving a couple of years late to this thread but I am encountering the same issue. Did you ever find the cause, or a workaround?

In my case I am processing a large file from a blob storage loading it into memory and then writing out the outputs to the blob storage again. With a small file I can see the full logs most of the time but sometimes the refresh button will still trigger a restart of the container. When dealing the full-size file, I will be able to see the logs for a bit and then systematically the refresh button will trigger a restart of the container. I did the test twice now of launching the ACI at night and not looking into the logs nor refreshing the status page and by the next day the process has finished and I have good results. However, I have no access to logs once the process has ended and it would be a pain to debug any issues. In addition to the fact that the ACI should restart at all by clicking on the refresh button.

I would very much appreciate any light you could shed on this. In my case:

Image: in ACR

Ports: 80

CPU cores: 4

Memory: 16 GiB

GPU SKU: None

GPU count: 0

OS type: Linux

Thank you in advance!

Felipe