Steps to reproduce:
1. On opencast01 QA server, schedule an individual or group recording for 5 minutes duration
2. Once it gets to the capture phase, it stays there and does not complete
I spoke with Adam about this, who thinks maybe the captures failed, but the CA didn't notify the workflow that it failed? You can check the capture screen on the opencast01 QA server to see 2 recordings that I scheduled that are stuck.
here is the additional log from the admin server (opencast01). http://pastebin.com/BEFZgteM
Apparently this prevents multi-ingest from working. I'm downgrading this but I think it would be a prime candidate for a maintenance release since we put so much time into satisfying this use case.
I put opencast03 ingest in maintenance mode and confirmed that the capture works when ingesting to admin server (no remote workflow service). Is it required to have the workflow server deployed with ingest? Can you have more than one workflow service deployed in a cluster?
Ruediger, I imagine now that ingest is working, we shouldn't find any more partially ingested files. Can you confirm? I suggest wiping out the ingest folder and scheduling a bunch of recordings to check. I also wonder whether this is the cause of "too many open file" errors.
I am changing this ticket's subject to reflect the main issue.
I get the following when trying update a workflow instance. Could be related?
Parameter Class org.opencastproject.workflow.api.WorkflowInstance has no constructor with single String parameter, static valueOf(String) or fromString(String) methods
Fixed the workflow rest endpoint's POST /update operation in r10253.