This just started happening today out of nowhere. I had a Galaxy Docker instance running fine for several days, but had to take it down. When I tried to start a new instance, Galaxy will not load. Docker reports no errors, but when I try to navigate to the Galaxy instance's URL, nothing loads and it says "Can't connect to server". I've tried several times, and this happens every time now. This is the command I am using to start a new Galaxy Docker instance:
docker run --name galaxy_test --cidfile=$PID_file -d -p 8080:80 -p 8021:21 -v $GALAXY_STORAGE_DIR:/export/ bgruening/galaxy-deeptools
I need to have a "production" and a "test" instance, so I gave it a name with the --name galaxy_test flag. I ran into this problem with duplicate names so I ran docker rm <name> on all the old container instances. Now, none of the new ones I make work. If I enter the container and run nginx, the browser page changes to a 403 Error, but thats it. Is there any documentation on how to actually run Galaxy from within the container? I am not sure what other programs within the container I need to look at for troubleshooting.
EDIT: As a followup, after waiting 20+ minutes, Galaxy is finally starting in these instances. Any ideas why it is taking so long when there was never a delay before?
You'd need to look in the logs for why there's such a delay. 20 minutes seems really excessive (our internal instance takes maybe 5 minutes and our public deepTools instance takes maybe 1 or 2). My guess is that there was another process on the server bound to port 8080, which caused nginx to keep crashing and supervisord to keep restarting it. After 20 minutes whatever had bound to that port must have been released, allowing nginx to start up. I've had that sort of issue before when I accidentally had apache running for no good reason.
Do you mean server logs, or logs within the container? Any idea where they would be located? I have been using 8080 as the port for all my Galaxy Docker instances but I have only ever had one running at a time.
Logs in the container, which are in
/home/galaxy/logs(well, the Galaxy-specific stuff is there).Please try to run your container with
-i -tand remove the-doption, this way you see a little bit of error/warning messages. Documentation about starting/restarting Galaxy is described here: https://github.com/bgruening/docker-galaxy-stable#Restarting-Galaxy We are using supervisord to control all demons. Moreover,-p 9002:9002will expose a webapp to start/stop processes for you.I've been using the
galaxy-deeptoolscontainer since it already has a few tools I want pre-installed. But it looks like port 9002 is not set up for that one? I added-p 9002:9002but the page doesn't load. https://github.com/bgruening/docker-recipes/blob/master/galaxy-deeptools/Dockerfile EDIT: Looks like I spoke too soon again, after 10 minutes it started working. Still looking into why its taking so long, none of the logs seem to have anythingIt's probably because we are copying many files. We are working on this to make it faster, it should be faster if you use /export and if you restart the container a second time.