<br><br><div class="gmail_quote">On Sat, Oct 22, 2011 at 12:01 AM, <span dir="ltr"><<a href="mailto:clownix@clownix.net">clownix@clownix.net</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
In my machines there is also sshd and other things that are usually there :<br></blockquote><div>hmm i do not expect ssh server to be ressource hungry</div><div><div><br></div><div>but a serial console is enough, imho</div>
</div><div>(or even nothing,.. (-;)</div><div><br></div><div>i had around 2-3MB ram usage per container (debian minimal rootfs) for a simple system (just init, getty, rsyslog, olsrd)</div><div>and 1.5MB for containers with just olsrd running,.. (lxc-init, olsrd)</div>
<div><br></div><div>(those numbers are without additional olsrd/system RAM usage for topology/routes)</div><div><br></div><div>so in both cases RAM (i had 16GB) was not my limit, (except for extremely dense topologies,..)</div>
<div><br></div><div>(but having only 2 cpu cores, was the limit)</div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<br>
[Clown1.1.1> ps -ef<br>
UID PID PPID C STIME TTY TIME CMD<br>
root 1 0 0 21:47 ? 00:00:00 init [2]<br>
root 253 1 0 21:47 ? 00:00:00 /usr/sbin/rsyslogd -c4<br>
root 264 1 0 21:47 ? 00:00:00 /usr/sbin/cron</blockquote><div>running cron , is asking for troubles </div><div>(as it might start something simultaneous on all systems)</div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
</blockquote><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
root 324 1 0 21:47 ? 00:00:00 dhclient -v -pf<br>
/var/run/dhclien<br></blockquote><div> </div><div>hmm who needs a dhcp cllient, when running ospf/olsrd? </div><div><br></div><div>except this, and that you are logged in (-;</div><div><br></div><div>your system looks nice && small aswell.</div>
<div><br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
root 361 1 0 21:47 ? 00:00:00 /usr/sbin/sshd<br>
root 373 1 0 21:48 ? 00:00:00 /usr/sbin/olsrd -d 0<br>
root 376 1 0 21:48 console 00:00:00 /bin/login -f console<br>
root 377 376 0 21:48 console 00:00:00 -bash<br>
root 380 361 0 21:48 ? 00:00:00 sshd: root@pts/0<br>
root 382 380 0 21:48 pts/0 00:00:00 -bash<br>
root 387 382 0 21:48 pts/0 00:00:00 ps -ef<br>
[Clown1.1.1><br>
<br>
May-be this explains the difference, and also may-be more machines can be<br>
handled with olsr. I had a big problem with ospf because when the topology<br>
is big ospf eats up too much memory and cpu per machine, so I did not push<br>
the olsr protocol much further.<br></blockquote><div>ok (-; </div><div><br></div><div>btw. i found that olsrds internal scheduler (of stable branch) uses too much cpu aswell, especially in such setups.</div><div>so infact on my tegra2 setup, 75% of the cpuload, was introduced by the schedulers.</div>
<div>(calculating the routes, parsing/generating messages, and so on, used only 20%)</div><div><br></div><div>btw my above numbers (250/2000 instances) are with olsrd 0.6.2 from stable branch with default config.</div><div>
with raising the PollRate, even more olsrd instances are possible on same cpu.</div><div>furthermore next major releases of olsrd will have better scheduling, too. </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<br>
The goal of cloonix is to manage the networking and the<br>
creation/destruction of the machines, not really to setup a record of<br>
machines.<br></blockquote><div>sure. </div><div><br></div><div>my inital idea was to evaluate if one can run/simulate a reasonable complex mesh-network on a "smartphone"</div><div>e.g. 250 olsrd instances, 350 interfaces, ~2000 links.</div>
<div>on a tegra2 system (cortex a9, 2x 1ghz, 512MB RAM)</div><div><br></div><div>so yes, it was a bit like record hunting (on limited hardware,..)</div><div><br></div><div>and afterwards i just checked how it would scale on a real PC.</div>
<div><br></div><div>and as your numbers of instances where (normalized by cpu performance) only around 1/15th of my setup,.</div><div>(which might indicate severe overhead somewhere in your setup)</div><div><br></div><div>
i just had to say something (-;</div><div><br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">per machine 2000 machines = 10 giga!<br></blockquote><div>btw i ran ~ 2000 on a dual core (5Ghz) system, </div>
<div><br></div><div>a empty machine itself should need quite nothing</div><div>(if its a linux container and not really a (virtual) machine)</div><div><br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
And also my file-systems are duplicated except for /bin /sbin /lib and<br>
/usr,</blockquote><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">and this takes some hard disk, 2000 would be far too much.<br></blockquote><div>(depends on the setup,..) </div>
<div><br></div><div>btw i ran my system on (nearly) complete readonly filesystems (-; </div><div>(i just had ~100KB harddisk usage per instance (-;)</div><div><br></div><div>or i just started a olsrd process inside a container (with its own network stack), <br>
</div><div><br></div><div>which imho for many testcases is just enough, as olsrd has its plugins that allow querying its state via network,..</div><div>and booting hundreds/thousands of systems (even if they aren`t RAM-hungry), just takes too much time,.. (-;</div>
<div><br></div><div>Markus</div></div>