FuzEc2 is a dumb helper script to automate all the stuff I was doing
manually for testing stacks of EC2 instances running a webapp on Fuzed.
Usage is thus:
$ cp fuze.example fuze
$ vim fuze (change all the configs)
$ ./fuze
This will list the things you can do with the script
$ ./fuze spin
This will show you all the instance types you can spin, eg:
available types:
INST TYPE RAM(GB) CU DISK ARCH COST/H
m1.small 1.7 1 160 32 0.1
m1.large 7.5 4 850 64 0.4
m1.xlarge 15.0 8 1690 64 0.8
c1.medium 1.7 5 350 32 0.2
c1.xlarge 7.0 20 1690 64 0.8
Then you can start spinning instances via:
$ ./fuze spin m1.small
To see which instances you have running, type:
$ ./fuze list
i-213ff448 domU-12-31-39-00-C0-02 ec2-75-101-131-181.compute-1.amazonaws.com m1.small running
i-1839f271 domU-12-31-39-00-5D-C6 ec2-67-202-36-240.compute-1.amazonaws.com m1.small running
i-a633f8cf domU-12-31-39-00-5D-87 ec2-75-101-225-2.compute-1.amazonaws.com m1.small running
i-b839f2d1 domU-12-31-39-00-E8-C3 ec2-75-101-131-128.compute-1.amazonaws.com m1.small running
i-5732f93e domU-12-31-39-00-8D-91 ec2-75-101-205-249.compute-1.amazonaws.com c1.xlarge running
i-0b33f862 domU-12-31-39-00-8D-71 ec2-75-101-200-57.compute-1.amazonaws.com c1.xlarge running
i-c833f8a1 domU-12-31-39-00-95-11 ec2-75-101-213-58.compute-1.amazonaws.com c1.xlarge running
i-163cf77f domU-12-31-39-00-22-21 ec2-67-202-40-157.compute-1.amazonaws.com m1.xlarge running
i-3b3df652 domU-12-31-39-00-C0-B7 ec2-75-101-254-200.compute-1.amazonaws.com m1.small running
To make those instances into certain Fuzed node types, do thusly:
$ ./fuze attach_master i-b839f2d1 # creates a master node - do this first
$ ./fuze attach_faceplate i-1839f271 # creates a faceplate
$ ./fuze attach_frontend i-213ff448 # creates an haproxy node - do this after you add faceplates
$ ./fuze attach_rails i-5732f93e # creates a rails node
Then you'll probably want to assign your frontend a static IP:
$ ec2-allocate-address # to get a static IP if you don't have one
$ ec2-associate-address YOUR.NEW.IP.ADDY -i i-213ff448 # assign your static IP to your haproxy node
To add more rails backend nodes, just do:
$ ./fuze spin c1.xlarge # wait a minute for it to spin
$ ./fuze list # grab the new i-xxxxxx instance id
$ ./fuze attach_rails i-xxxxxx
That's it - your capacity goes up. For faceplates, you do this:
$ ./fuze spin m1.small # wait a minute for it to spin
$ ./fuze list # grab the new i-xxxxxx instance id
$ ./fuze attach_faceplate i-xxxxxx
$ ./fuze attach_frontend i-213ff448 # re-run this to make haproxy rewrite and restart
You can see the stats on your haproxy node by visiting here:
http://your-frontend-address/haproxy?stats
And then enter the username/password you put in your config file to get access.
Here is an example of one big-ass cluster:
$ ./fuze list
frontend i-213ff448 domU-12-31-39-00-C0-02 ec2-75-101-131-181.compute-1.amazonaws.com m1.small running
faceplate i-1839f271 domU-12-31-39-00-5D-C6 ec2-67-202-36-240.compute-1.amazonaws.com m1.small running
faceplate i-a633f8cf domU-12-31-39-00-5D-87 ec2-75-101-225-2.compute-1.amazonaws.com m1.small running
faceplate i-4a33f823 domU-12-31-39-00-5D-D4 ec2-67-202-52-242.compute-1.amazonaws.com m1.small running
master i-b839f2d1 domU-12-31-39-00-E8-C3 ec2-75-101-131-128.compute-1.amazonaws.com m1.small running
rails i-5732f93e domU-12-31-39-00-8D-91 ec2-75-101-205-249.compute-1.amazonaws.com c1.xlarge running
rails i-0b33f862 domU-12-31-39-00-8D-71 ec2-75-101-200-57.compute-1.amazonaws.com c1.xlarge running
rails i-c833f8a1 domU-12-31-39-00-95-11 ec2-75-101-213-58.compute-1.amazonaws.com c1.xlarge running
rails i-4f33f826 domU-12-31-39-00-90-21 ec2-75-101-199-42.compute-1.amazonaws.com c1.xlarge running
rails i-0933f860 domU-12-31-39-00-8D-B1 ec2-75-101-206-32.compute-1.amazonaws.com c1.xlarge running
rails i-cd33f8a4 domU-12-31-39-00-91-61 ec2-75-101-205-227.compute-1.amazonaws.com c1.xlarge running
rails i-3633f85f domU-12-31-39-00-8C-71 ec2-75-101-195-69.compute-1.amazonaws.com c1.xlarge running
rails i-163cf77f domU-12-31-39-00-22-21 ec2-67-202-40-157.compute-1.amazonaws.com m1.xlarge running
rails i-3b3df652 domU-12-31-39-00-C0-B7 ec2-75-101-254-200.compute-1.amazonaws.com m1.small running
Watch out, though - that particular cluster costs about $6.40/hour to run
- thats $150 a day, so don't forget about it. It's unlikely you'll need it,
even with really crappy Rails apps, that should process a freak-load of requests
per second, it represents about 350 Rails handlers and something like 60 cores.