TLFox wrote:
>
> Are there any workarounds here?
>
>
In case you're interested, I have managed to work around this to a certain
level by manually throttling the producer on the client code that calls the
producer. I have a nice class that will automatically throttle given a
specified rate - it's very accurate based on the token bucket limiter
algorithm.
I started the throttle rate low (1000 msgs/sec) and gradually increased it
until qpid got into a "non steady state" (i.e. one that would eventually
lead to OOM), where the consumer can no longer consume messages at the rate
the producer produces.
This was about 5000 msgs / sec IIRC on my client and server setup (pretty
basic machines) (1K messages), using M2.1, non persistent
--
View this message in context:
http://www.nabble.com/Persistent-messages-sent-non-blocking-tp17517786p17537885.html
Sent from the Qpid Developers mailing list archive at Nabble.com.
