<p dir="ltr">I have only bad experiences with Azure, on all levels, from restarting machines without prior notice, problems with subscription that lead to downtimes and communication in the worst corporate style (tons of buzzwords, no content). </p>
<p dir="ltr">If you want to have experience like with virtual data center but use real one, pick OVH. They are automated to such extent that you quickly forget you operate on bare metal. But you have bare metal performance and prices. </p>
<p dir="ltr">M. </p>
<div class="gmail_quote">10.06.2016 1:00 PM "Popov, Zahar" <<a href="mailto:zahar.popov1978@yandex.com">zahar.popov1978@yandex.com</a>> napisał(a):<br type="attribution"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Hi Philipp<br>
Thank you for chiming in.<br>
<br>
The only reason i use the -kh fork is because it seemed to be more recent. I had this performance issue with the regular version and tried the fork.<br>
<br>
I realize that the problem can be with the TCP stack parameters. I had no problem getting about 800Mbps between these machines when using iperf. Certainly, the workload completely different, but at least i know that the TCP stack is somewhat operational.<br>
<br>
Do you happen to know specifically what’s broken?<br>
<br>
I don’t have access to a physical data center this is why i would like to use either EC2 or Azure.<br>
<br>
thanks!<br>
—zahar<br>
<br>
> On Jun 10, 2016, at 7:09 PM, Philipp Schafft <<a href="mailto:lion@lion.leolix.org">lion@lion.leolix.org</a>> wrote:<br>
><br>
> Good noon,<br>
><br>
> On Fri, 2016-06-10 at 12:50 +0900, Zahar Popov wrote:<br>
>> Hello<br>
>> I'm trying to measure the performance of the icecast relay server on<br>
>> 64kbps streams.<br>
>><br>
>> The server is running in AWS [...]. I'm using the icecast-kh fork.<br>
>> [...]<br>
>> I'm able to go up to around 9K simultaneous connections to the server<br>
>> (from two machines). [...]<br>
><br>
> First: -kh is a independent fork. So I can hardly speak for it. This<br>
> answer is completely based on the (often wrong) assumption that in this<br>
> case behaves exactly like the official Icecast2 in any stable and<br>
> supported version.<br>
><br>
><br>
>> It doesn't matter if i run one or more instances of the relay server,<br>
>> the limit seems to be OS global so when one instance is running with<br>
>> 5K connections and the other instance is getting close to 4K<br>
>> connections they both start dropping connections.<br>
>><br>
>> I assume that there is some other setting of the stack that i didn't<br>
>> configure so i was wondering if anybody was able to run a few dozens<br>
>> of thousands of connections on one server.<br>
><br>
> This should give you the hint: It's not a problem of Icecast2 as it<br>
> doesn't depend on the processes.<br>
><br>
> Basically: AWS is well known for their broken TCP stack. There is no<br>
> reason to run Icecast2 on AWS. Just use a normal server and be fine.<br>
> What you see above is just AWS's maximum connection limit per instance<br>
> or something. (But there are more problems known with AWS's TCP stack<br>
> that affect real world deployments of Icecast2.)<br>
><br>
><br>
> Have a nice day!<br>
><br>
> --<br>
> Philipp.<br>
> (Rah of PH2)<br>
<br>
_______________________________________________<br>
Icecast-dev mailing list<br>
<a href="mailto:Icecast-dev@xiph.org">Icecast-dev@xiph.org</a><br>
<a href="http://lists.xiph.org/mailman/listinfo/icecast-dev" rel="noreferrer" target="_blank">http://lists.xiph.org/mailman/listinfo/icecast-dev</a><br>
</blockquote></div>