<div dir="ltr">Makes sense. <div>I'm trying to reproduce with higher watermark by publishing messages, consuming them but not acking (this is what happened in production) but can't get the memory to go above the watermark.</div>
<div>Can you explain or point me to an explanation of Rabbit's memory usage?</div><div><br></div><div>Also, When hitting the watermark, the server seems to block the producer on the socket level, instead of sending a flow control message. Since the .Net SDK doesn't expose a send timeout, publishing is blocked forever.</div>
<div>Any reason for not exposing such timeout?<br><div><br clear="all"><div dir="ltr"><div><br></div><div>Thanks<br clear="all"><div dir="ltr"><i style="color:rgb(102,102,102)">--</i><span style="color:rgb(102,102,102)">Raviv</span></div>
</div><div><font color="#666666"><i><br></i></font></div></div><br>
<br><br><div class="gmail_quote">On Thu, Oct 4, 2012 at 6:17 PM, Simon MacMullen <span dir="ltr"><<a href="mailto:simon@rabbitmq.com" target="_blank">simon@rabbitmq.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div class="im">On 04/10/12 16:57, Raviv Pavel wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Back to the original problem:<br>
1. I set the watermark very low so I can reach it, effectively 78MB<br>
</blockquote>
<br></div>
Ah. Was this what you had done in the first place?<div class="im"><br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
2. server starts with empty queues and memory usage is 66MB<br>
3. I start publishing messages and reach the watermark<br>
4. publisher gets blocked<br>
5. I purge the queue but memory stays the same<br>
</blockquote>
<br></div>
When you set the watermark very low, close to the "empty" memory use, it is very possible that the memory level will go over and never go under again. Caches, lazy instantiation and so on. RabbitMQ is not designed for highly constrained memory environments.<br>
<br>
Cheers. Simon<div class="HOEnZb"><div class="h5"><br>
<br>
-- <br>
Simon MacMullen<br>
RabbitMQ, VMware<br>
</div></div></blockquote></div><br></div></div></div>