ClusterODM Not respecting --max-images flag?

As smathermather-cm wrote elsewhere that it’s important to set the --max-images flag arbitrarily high on the primary machine.

Primary machine NodeODM setup like
docker run -p 3001:3000 opendronemap/nodeodm -q 2 --max-images 1000000

Rest of cluster NodeODM setup like
docker run -p 3000:3000 opendronemap/nodeodm -q 2 --max-images 200

However, my secondary machines are getting job sizes of >400 images. Also the “main” named Task of 1123 images usually shows up on the localhost:3000 page of a secondary machine.

I understand that --max-images is an average, but I’m mostly concerned about the main task showing up on less powerful machines. Cluster keeps failing due to running out of RAM, so trying to nail down the issue.

This smells like user error, but I’ve reset each machine with the parameters listed above.

Lowest RAM on a secondary machine is 64gb. Thank you!

--max_images not --max-images (note _ vs -).

We probably should improve that (or print an error if it’s the wrong syntax).

2 Likes

My hunch was right! User error! :grimacing: :sweat_smile:

I’ll go slam my forehead into my desk now… Probably never would have caught that. Thank you.

2 Likes

That’s an error I have made a lot.

1 Like

I’ve made that error too… and I wrote it lol

3 Likes

What are everyone’s thoughts about aligning the usage of _ in these parameters to - like everything else?

1 Like
4 Likes