Using progress=True results in hung process after a previous run of sync process #121
Assignees
Labels
bug
Something isn't working
Comments
tlvu
added a commit
to bird-house/birdhouse-deploy
that referenced
this issue
Apr 15, 2020
Add generic_bird optional component. Could be used as work-around for issue bird-house/finch#121 so we have a Finch dedicated for async request only. This component is generic and configurable enough it could even accommodate any WPS. It has its own Postgres instance so we can use it later to experiment with different version of Postgres. Test server: https://lvupavics-lvu.pagekite.me/twitcher/ows/proxy/finchasync, https://lvupavics-lvu.pagekite.me/canarie/node/service/status See the `README.md` update for more info. Note the docker image, service name and port is customizable (my test server do not use the default service name). It starts out as a Finch2 but now its rather a generic skeleton to plugin any birds we need. I wish I had this when I was testing out the new Thunderbird from PCIC. At some point in the past, we had the idea of 2 simultaneous FlyingPigeon because the new one is so different we wanted to keep the old one around for a short while. This PR could have helped. So do not view this PR as Finch 2 anymore, view this PR as a mechanism to quickly deploy another bird, any bird, without having to change code. This opens up many options, testing is just one of the possible usage.
|
Upcoming "jobqueue" feature from PyWPS (geopython/pywps#505) might help fix this. |
|
Running without dask resolves this problem. |
|
@eyvorchuk seems to have hit this problem again so the fix might not have worked for all cases. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Description
Running the following notebook, the async process never terminate
finch_test-pavics.ipynb.txt
But restarting the Finch server (
pavics-compose.sh restart finch) and clear the queue left by the previous hung process (scripts/clear-running-wps-jobs-in-db.sh finch $POSTGRES_PAVICS_USERNAME), to avoid the max connection problem, then only run the async process (do not run the sync process before) and it works.$POSTGRES_PAVICS_USERNAMEis fromenv.localfile.Note it's restarting the WPS server (Finch) that unblock async request until the next sync request and async will be blocked again. Clearing the queue is just a nice to have since old items will eventually burst the max limit of concurrent item in the queue.
Environment
Additional Information
Might relate to #45
Edit:
The text was updated successfully, but these errors were encountered: