After installing the latest webODM through docker and messing around with it a bit, I now have a queue count of 13 on node-odm-1, and new tasks won’t run (status queued).
Access the web interface of the node directly shows 0 running tasks.
How do I clear the list? (or alternatively, how to start over, preferably without having docker download all the images again?)
Duh… For future googlers/me:
How can the queue be reset on a native install too? Jobs seem to get ‘stuck’; they aren’t processing and webodm isn’t aware of them (other than the queue count from nodeodm).
So the job is stuck in the nodeodm instance? At a guess, I’d try restarting the nodeodm job, but I’ve not seen that particular issue I’m afraid.
That’s the thing, there’s no job to restart. It’s like the counter in the queue doesn’t get updated sometimes when a job finishes or dies. I’ve tried a reinstall/upgrade of the node but the issue remains. I’ve been through the API docs but there is no option to purge the queue (only display its size). Short of a complete system wipe, I don’t know what to try.
The only time I’ve seen NodeODM hang and fail to update the queue is with this scenario. https://github.com/OpenDroneMap/ODM/issues/944 is that what you are seeing?
To reset the queue, just remove the contents of the
data/ directory (folders and .json file) and restart NodeODM.
Possibly, I was having pdal hanging issues - it seems to be extremely sensitive to the GCPs that it’s fed. I didn’t dig too deeply into each fault because I figured that I was using it wrong rather than it being a bug.
I managed to get the queue to zero by a combination of upgrading and rebooting - though I’ve tried both individually before and it hadn’t helped. I’ll keep the data/ trick in mind, thanks.
Forgive my ignorance but where is the data/ directory folder you are referring to?
I’m using WebODM and I seem to be having the same issue. My computer went to sleep while processing a large dataset, now it seems that there are 2 jobs in the queue, the processor is working super hard, but I don’t see anywhere in WebODM that the status of a job is being worked on. If I try to start anything it adds it to the queue.
For native installs it’s /www/data
I can’t remember where it was on Docker, but if you run a find for the file “tasks.json” it should show up.
The problem is caused by a container of docker who stucked.
docker container ls, then look the container ID who got stuck.
After that you can kill it by
docker container kill container_ID.
Restart your node ODM / ODM
I know I am reopening a old topic but it happend me right today, and the best solution a consider is acces to node from web socket where task are presented by UUID and you can cancel, restart or delete.
Hey, I’m pretty new to this and I also have the same issue. Where can I run this code?
For anybody looking for a solution for the native version of WebODM what you need to do is:
Close out of the WebODM application
Locate your install folder (Mine was in the C: drive)
Then go through the following folders:
WebODM > resources > app > apps > NodeODM > data
Once you have reached the data folder delete all of the contents
After starting WebODM you will see that the current tasks have failed and you just need to restart the tasks. Your tasks should begin processing at this point.