[
http://opencast.jira.com/browse/MH-8092?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=28297#comment-28297
]
Tobias Wunden commented on MH-8092:
-----------------------------------
Some more comments on this ticket: First of all, the ingest service immediately
creates a job for the ingest operation, so it should be easy enough to ask the
service registry for running jobs of type IngestServiceImpl.JOB_TYPE to get the
number of active ingest operations. The job will include all the information
that's needed, especially the host name, duration etc.
With that knowledge in mind, it is easy enough to implement a limitation. Just
ask the service registry for running ingest operations on the current host, and
if that exceeds x, return SERVICE_UNAVAILABLE (503). I am wondering if we
should use the service registry's knowledge on the maximum number of parallel
jobs on that machine as a limit rather than a configuration value.
The downside of this approach is that the capture agent will need to be able to
deal with this and come up with a good strategy around it.
> Limit number of ingests
> -----------------------
>
> Key: MH-8092
> URL: http://opencast.jira.com/browse/MH-8092
> Project: Matterhorn Project
> Issue Type: Task
> Components: Administrative Tools
> Affects Versions: 1.1 , 1.2
> Reporter: Christopher Brooks
> Assignee: Adam McKenzie
> Priority: Blocker
> Fix For: 1.3
>
>
> A flag to the ingest service should be added to allow it to limit itself to X
> number of simultanious ingests at a time. This flag should be set through a
> configuration property. The implementation should use a static collection
> which lists all of the currently running ingests by hostname, so that the
> admin page could include text like "There are current 2 ingests:
> thorv105.usask.ca, thorv205a.usask.ca".
> When ingests fail or complete the associated dns entry should be removed from
> the collect. When ingests are initiated the collection should be syncronized
> upon and the dns entry added if it is not at the limited size. If the
> collection is at max size, an http 503 code of service unavailable should be
> sent back.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
http://opencast.jira.com/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
_______________________________________________
Matterhorn mailing list
[email protected]
http://lists.opencastproject.org/mailman/listinfo/matterhorn
To unsubscribe please email
[email protected]
_______________________________________________
