[zeromq-dev] Handling OOM

[Ilja Golshtein]

18.05.2011, 17:58, "Pieter Hintjens" <ph at imatix.com>:
> My personal preference for handling OOM is to die, rapidly, with a
> message like "Raise your VM size and restart everything!". 

Same thing.

I've already written what I think about this
http://lists.zeromq.org/pipermail/zeromq-dev/2011-May/011342.html

Although it is good idea to extend HWM stuff
http://lists.zeromq.org/pipermail/zeromq-dev/2011-May/011374.html

> But clearly
> asking to allocate a huge message can fail without signalling that
> other stuff will also fail.

In my own project there is at least one link 1M messages are common for.
It might be difficult to understand what is going on if some messages are silently disappear.

>From my point of view it is fine to trigger behavior based on message size if it is possible to
specify this threshold explicitly.

Thanks.

-- 
Best regards,
Ilja Golshtein.