Conversation:
Notices
-
@takeshitakenji I think the reason my queue daemons when I tried them segfault is because I have such a large queue at a given time, whereas in your case, they're fine, because your usual floating amount is apparently around 10.
-
@takeshitakenji It isn't so much the number per se. A large instance can be expected to have a large queue. The problem is that failed entries are staying in the queue, and every time the queue executes, be it on OpprotunisticQM or the daemons (STOMP is different though), all the failed entries are tried before the new ones. This is why Wrongthink ended up so fa…
-
@takeshitakenji Something about the queue daemons seems to fail on larger queues irrespective of whether the entries are new or old, though. Whereas OpprotunisticQM is "ok" - it "just" slows down when this happens (potentially a lot)
-
@maiyannah @takeshitakenji on Freedombone as a workaround I have a script which looks for the presence of those daemons and restarts them if they don't appear.
-
@bob @takeshitakenji I just use opprotunisticQM since I can afford the performance overhead, but I can understand not everyone can.
-
@takeshitakenji Its basically trading a memory overhead for greater performance/responsiveness, which is a tradeoff I can make since I'm on a dedicated machine, but people on smaller VPSes or running on hardware like RPis or beaglebones can't. Different use cases.
-
@moonman @takeshitakenji STOMP really is the best of ways but it requires setting up, operating, and maintaining that separate server.
-
@takeshitakenji @bob I've never had any issues with being behind, but again, that's because OQM makes the system specs of the underlying hardware much more important. If you have the performance it fares better. If you don't, daemons do.
-
@moonman @takeshitakenji Didn't that get superceded by Apollo?
-
@takeshitakenji @moonman http://activemq.apache.org/apollo/ http://stomp.github.io/stomp-specification-1.2.html
-
@takeshitakenji @moonman CONFIGURE explains the setup options but the GS end is no more complicated than setting up a database server. Point it to a URL with hostname port, supply user/pass, done. The more "complicated" thing is setting up the MQ server.
-
@takeshitakenji @moonman Yeah. The Apache Artemis or Apollo server are kind of hit or miss. Either they plug and play or you'll be configuring them for hours to get them to work. I've had both experiences in the past.
I'm just saying the GS\PA end is just dropping the credentials in config.php and enabling STOMP as the queue. Everything else on that side of things is handled in the back-end.
-
@moonman @takeshitakenji It's almost like it's an Apache product or something :^)
-
@moonman @takeshitakenji No, you need to move to a S3 compatible service that isn't run by Dick Dastardly like Amazon is.
-
@moonman @takeshitakenji I wonder how hard it would be to make a service that automatically deployed postActiv installs with a corresponding STOMP server.
I may actually look into this for serious after the render stuff is put to bed.
-
@takeshitakenji @moonman If it doesn't crash your server occasionally, it's probably either made by the Kernel group, or not a FOSS project.
-
@takeshitakenji @moonman It depends on what server you go with and what environment its going into but Moon's experience with this is probably more recent than mine since I set up the Artemis server I'm using for testing literally three years ago.
maiyannah
Bob Mottram
All Hacker Poesy content and data are available under the