[rabbitmq-discuss] server crashes with very fast consumers

[alex chen]

John,

> In many of my tests, I ran with a 32-bit Erlang, and with a relatively small 
>amount of RAM. RabbitMQ had to do quite a lot more work to conserve RAM, and 
>sometimes it failed to do so fast enough, and Erlang couldn't allocate more 
>memory, and crashed. 
>

that is exactly what happened in our tests.  the broker's memory keeps on 
increasing until it reaches near 4GB (32 bit limit), then it crashes.  i think 
if you increase the RAM, cpu cores (we use 8 core) and disk speed (RAID), the 
broker is more likely to crash, because the memory leak increases as messagne 
consuming rate increases.  you can move the consumers to other machines. as long 
as they are close to the broker, the consuming rate will still be very high, and 
broker would crash.  you mentioned that your machine started swaping while 
running the tests.  this swaping actually could slow down the messages reads 
from disk, so broker is less likely to crash.

thanks.

-alex
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20110414/6bff665f/attachment-0001.htm>