Changes

== Full VE container lifecycle ==

Create a VEcontainer, set an IP, start, add a user, enter, exec, show <code>ps -axf</code> output inside the VEcontainer, stop, and destroy. It should take about two minutes (''"compare that to a time you need to deploy a new (non-virtual) server!"''). During the demonstration, describe what's happening and why.

# VECT=123

# time vzctl create $VE CT --ostemplate fedora-core-5-i386-default # vzctl set $VE CT --ipadd $IP --hostname newVE newCT --save # vzctl start $VECT # vzctl exec $VE CT ps axf # vzctl set $VE CT --userpasswd guest:secret --save

[newVEnewCT]# ps axf [newVEnewCT]# logout # vzctl stop $VECT # vzctl destroy $VECT

== Massive VE container creation ==

Create/start 50 or 100 VEs containers in a shell loop. Shows fast deployment and high density.

# time for ((VECT=200; VECT<250; VECT++)); do \> time vzctl create $VE CT --ostemplate fedora-core-59-i386-default; \> vzctl start $VECT; \

== Massive VE container load ==

Use VEs containers from the previous item — load those by <code>ab</code> or <code>http_load</code>. This demo shows that multiple VEs containers are working just fine, with low response time etc.

# for ((VECT=200; VECT<250; VECT++)); do \> vzctl set $VE CT --ipadd 10.1.1.$VE CT --save; \

If you have two boxes, do "<code>vzmigrate --online</code>" from one box to another. You can use, say, <code>xvnc</code> in a VE container and <code>vncclient</code> to connect to it, then run <code>xscreensaver-demo</code> , choose a suitable screensaver (eye-candy butnot too CPU aggressive) and while the picture is moving do start a live migration. You'll show see that <code>xscreensaver</code> stalls for a few seconds but then keeps running continues to run — on another machine! That looks amazing, to say at least.

FIXME: commands, setup, vnc VNC template.

 We can see that the number of processes inside VE container will not be grow upgrowing. We will see only the increase of <code>numproc</code> and/or <code>kmemsize</code> fail counters in <code>/proc/user_beancounters</code>.

 {{Warning|CPU weights only works in stable kernels.}} Create 3 VEscontainers:

Set VEs container weights:

We set next cpu CPU sharing <code>VE101 CT101 : VE102 CT102 : VE103 CT103 = 1 : 2 : 3</code>

Run VEsStart containers:

Run busy loops in VEsall containers:

So, we see that CPU time is given to VEs container in proportion ~ 1 : 2 : 3. Now start some more busy loops. CPU distribution should remain the same.

# vzctl set VEID CTID --diskspace 1048576:1153434 --save# vzctl start VEIDCTID# vzctl enter VEIDCTID