I just started using the Windows version of WebODM.
I would like to process data on multiple machines simultaneously.
Is there a step-by-step guide to doing this?
I used this batch file I found to start a node on the remote machine:
cd resources\app\apps\NodeODM
nodeodm.exe --odm_path …\ODM -p 3000
I can connect to this, and I can assign jobs to the remote node or the local node, but I can’t figure out how to assign jobs to multiple nodes.
OK! So I think I am making progress, but still not there!
What I have done so far is -
I started native Windows nodeodm.exe on 3 machines, one with port 3001, because it looked like it interfered with clusterodm.exe running on same machine.
I started native Windows clusterodm.exe on the one machine.
I used Putty to telnet to the clusterodm server, and added the 3 machines.
I added the clusterodm server to my Windows native Webodm under PROCESSING NODES.
I think its OK to this point. Now when I try to run a small test I only see this:
Were you able to make it work? I have a couple of CPUs servers that I wouldn’t mind putting them back up to use for my personal WebODM. This will help load balance the build.
Hi, I was able to make it work. Couple of things to consider when setting the ClusterODM server in distrubted splite merge.
Ports to the ClusterODM API needs to be enabled. By default I believe is TCP 10000
When setting up your “options” in your project tast, you need fill these options:
sm-cluster (string)
split (positive integer)
split-overlap (positive integer)
This will give your ClusterODM enough information to assign a master nodeodm. master nodeodm will call back to clusterodm with split tasks assignment requests. master nodeodm does the merge.