Thursday, December 13, 2012

[SWIG] How to add C function in Python by using SWIG

Here is a simple example to use SWIG to automatically wrap C function and generate a wrapper and build a shared library for python.

/*** File : example.c ***/
#include 
<time.h>
double My_variable = 3.0;
int fact(int n) {
if (n <= 1) return 1;
else return n*fact(n-1);
}
int my_mod(int x, int y) {
return (x%y);
}
char *get_time()
{
time_t ltime;
time(&ltime);
return ctime(&ltime);
}
/*************************/


/*** example.i ***/
%module example
%{
/* Put header files here or function declarations like below */
extern double My_variable;
extern int fact(int n);
extern int my_mod(int x, int y);
extern char *get_time();
%}
extern double My_variable;
extern int fact(int n);
extern int my_mod(int x, int y);
extern char *get_time();
/******************/
Another example:
/* File : example.c */

double  My_variable  = 3.0;

/* Compute factorial of n */
int fact(int n) {
  if (n <= 1)
    return 1;
  else
    return n*fact(n-1);
}

/* Compute n mod m */
int my_mod(int n, int m) {
  return(n % m);
}

/* File : example.i */
%module example
%{
/* Put headers and other declarations here */
extern double My_variable;
extern int    fact(int);
extern int    my_mod(int n, int m);
%}

extern double My_variable;
extern int    fact(int);
extern int    my_mod(int n, int m);

What if you have some pointer arguments in C ? How to deal with it in Python? 

For example:void add(double a, double b, double *result) {
 *result = a + b;
}
Please refer to this document. It will give you the nametype mapping in SWIG.http://www.swig.org/Doc1.3/Arguments.html





Wednesday, December 12, 2012

[Trema] The L2 isolation mechanism in sliceable switch

If someone has ever seen the documents about sliceable switch as below, he/she will feel headache or sick because of a lot of contents and description.
https://github.com/trema/apps/wiki/sliceable_switch_tutorial
https://github.com/trema/apps/wiki/sliceable_switch_features

Now, I will give a flow control chart of slice function which is summarized from the source code ( slice.c ). That can give you a clear image about L2 isolation mechanism in sliceable switch, specially in Slice function. Check it out as the following chart:

 So, broadly speaking, the slice function will check mac binding first, then port_mac binding, and finally port binding. Meanwhile, some configurations will affect the result, for instance, "restrict hosts on port" enabled will force the slice function to check port_mac binding, otherwise, it won't do that.
Based on this flow chart, you can compare with the test cases in https://github.com/trema/apps/wiki/sliceable_switch_features







Monday, December 10, 2012

[MongoDB] Install MongoDB and try a simple example of mongodb_c_driver

MongoDB Installation
http://docs.mongodb.org/manual/tutorial/install-mongodb-on-debian-or-ubuntu-linux/
or
https://www.digitalocean.com/community/articles/how-to-install-mongodb-on-ubuntu-12-04
For instance in my environment:
  > sudo apt-key adv --keyserver keyserver.ubuntu.com --recv 7F0CEB10
  > sudo echo "deb http://downloads-distro.mongodb.org/repo/ubuntu-upstart dist 10gen" | tee -a /etc/apt/sources.list.d/10gen.list
  > sudo apt-get -y update
  > sudo apt-get -y install mongodb-10gen

Try command on MongoDB
>mongo
MongoDB shell version: 2.2.2
connecting to: test

> show dbs
db    (empty)
local    (empty)
test    0.203125GB

 > use test
switched to db test

> show collections
foo
system.indexes

### Insert new data ###
> db.foo.save({a:1})
> doc = {
... "name" : "kristina",
... "contact info" : {
... "twitter" : "@kchodorow",
... "email" : "kristina@10gen.com"
... },
... "friends" : 400232,
... "pic" : BinData(...)
... "member since" : new Date()}
> db.foo.insert(doc)
> db.foo.save({a:2})
> db.foo.save({a:3})

### Find the data ###
> db.foo.find()
{ "_id" : ObjectId("50c592540770027c182d31b9"), "a" : 1 }
{ "_id" : ObjectId("50c59e030770027c182d31ba"), "name" : "kristina", "contact info" : { "twitter" : "@kchodorow", "email" : "kristina@10gen.com" }, "friends" : 400232, "member since" : ISODate("2012-12-10T08:30:50.389Z") }
{ "_id" : ObjectId("50c59e1f0770027c182d31bb"), "a" : 2 }
{ "_id" : ObjectId("50c59e220770027c182d31bc"), "a" : 3 }

> db.foo.findOne()
{ "_id" : ObjectId("50c592540770027c182d31b9"), "a" : 1 }
> db.foo.find({"a":1})
{ "_id" : ObjectId("50c592540770027c182d31b9"), "a" : 1 }
> db.foo.find({"name":"kristina"})
{ "_id" : ObjectId("50c59e030770027c182d31ba"), "name" : "kristina", "contact info" : { "twitter" : "@kchodorow", "email" : "kristina@10gen.com" }, "friends" : 400232, "member since" : ISODate("2012-12-10T08:30:50.389Z") }

Try a simple example of mongodb c driver
For more API info in details, please refer to this: http://api.mongodb.org/c/current/tutorial.htm
 

> gcc --std=c99 -I/usr/local/include -L/usr/local/lib -o mongodb_test mongodb_test.c -lmongoc
> ./mongodb_test
WARNING: mongo_connect() is deprecated, please use mongo_client()
MONGO_OK:connection succeeded
0


mongodb_test.c (source code)

#include <stdio.h> #include "mongo.h" int main() { mongo conn[1]; int status; status = mongo_connect( conn, "127.0.0.1", 27017 ); if( status != MONGO_OK ) { switch ( conn->err ) { case MONGO_CONN_SUCCESS: printf( "connection succeeded\n" ); break; //case MONGO_CONN_BAD_ARG: printf( "bad arguments\n" ); return 1; case MONGO_CONN_NO_SOCKET: printf( "no socket\n" ); return 1; case MONGO_CONN_FAIL: printf( "connection failed\n" ); return 1; case MONGO_CONN_NOT_MASTER: printf( "not master\n" ); return 1; } }else{ printf( "MONGO_OK:connection succeeded\n%d\n", status ); } mongo_destroy( conn ); return 0; }


[Memcached] Install memcached and try libmemcached C API


Install from package
> sudo apt-get install memcached

or Install from source code
We need to have :
  1. libevent downloaded from : http://libevent.org/ 
    • > ./configure --prefix=/usr
    • > make
    • > sudo make install
  2. memcached downloaded from : http://memcached.org/
    • > ./configure --prefix=/usr/local
    • > make
    • > sudo make install
Check the status of memcached
 > sudo service memcached status


Install libmemcached C API from source code

  1. libmemcached C API ownloaded from http://libmemcached.org/libMemcached.html
    • > ./configure --prefix=/usr
    • > make 
    • > sudo make install

Give a simple try for libmemcached C API
> gcc -o mem_test2 mem_test2.c -lmemcached -lpthread
> ./mem_test2
Save key:key1 data:"This is c first value" success.
Fetch key:key1 data:This is c first value
Delete Key key1 success.

mem_test2.c

#include <stdio.h> #include <stdlib.h> #include <string.h> #include <libmemcached/memcached.h> int main(int argc, char *argv[]) { memcached_st *memc; memcached_return rc; memcached_server_st *servers; char value[8191]; //connect multi server memc = memcached_create(NULL); servers = memcached_server_list_append(NULL, "localhost", 11211, &rc); servers = memcached_server_list_append(servers, "localhost", 11212, &rc); rc = memcached_server_push(memc, servers); memcached_server_free(servers); //Save multi data size_t i; char *keys[] = {"key1", "key2", "key3"}; size_t key_length[] = {4, 4, 4}; char *values[] = {"This is c first value", "This is c second value", "This is c third value"}; size_t val_length[] = {21, 22, 21}; for (i = 0; i < 3; i++) { rc = memcached_set(memc, keys[i], key_length[i], values[i], val_length[i], (time_t) 180, (uint32_t) 0); if (rc == MEMCACHED_SUCCESS) { printf("Save key:%s data:\"%s\" success.\n", keys[i], values[i]); } } //Fetch multi data char return_key[MEMCACHED_MAX_KEY]; size_t return_key_length; char *return_value; size_t return_value_length; uint32_t flags; rc = memcached_mget(memc, keys, key_length, 3); while ((return_value = memcached_fetch(memc, return_key, &return_key_length, &return_value_length, &flags, &rc))) { if (rc == MEMCACHED_SUCCESS) { printf("Fetch key:%s data:%s\n", return_key, return_value); } } //Delete multi data for (i = 0; i < 3; i++) { rc = memcached_set(memc, keys[i], key_length[i], values[i], val_length[i], (time_t) 180, (uint32_t) 0); rc = memcached_delete(memc, keys[i], key_length[i], (time_t) 0); if (rc == MEMCACHED_SUCCESS) { printf("Delete %s success\n", keys[i], values[i]); } } //free memcached_free(memc); return 0; }



Wednesday, December 5, 2012

[Presentation] OpenStack 2012 fall summit observation - Quantum/SDN

Taiwan OpenStack User Group (TWOSUG)3rd Meet Up is hold in Dec 5, 2012. I give a presentation in one of session, which is "OpenStack 2012 fall summit observation - Quantum/SDN". The topic is focused on Quantum and SDN and the slide is shared on shlideshare as follows:
http://www.slideshare.net/teyenliu/open-stack-2012-fall-summit-observation-with-quantumsdn-15493510

Friday, November 23, 2012

[iptables] some common examples of iptables rule


  • Read all tables without DNS lookup
    • > iptables -L -n
  • Obtain the line number of the lines: 
    •  > iptables -L -nv --line-numbers
  • Read NAT table in list without DNS lookup
    • > iptables -t nat -L -n
  • Do NAT ( SNAT )
    • > echo "1" > /proc/sys/net/ipv4/ip_forward
    • > iptables-t nat -A POSTROUTING -s ${INSIDE_NETWORK}/${INSIDE_NETMASK} -o ${OUTSIDE_DEVICE} -j MASQUERADE
    • or > iptables-t nat -A POSTROUTING -s ${INSIDE_NETWORK}/${INSIDE_NETMASK} -o ${OUTSIDE_DEVICE} -j SNAT --to ${TARGET_IP}
  • Do DNAT 
    • > iptables -t nat -A PREROUTING -i eth0 -p tcp --dport 80 -j DNAT --to-destination 192.168.100.10:80
  • Drop the packet which is from 192.168.2.20 to 192.168.1.100 with TCP port 80
    • > iptables -A POSTROUTING -t nat -s 192.168.2.20 -d 192.168.1.100 -p TCP --dport 80 -j DROP
  • Accept the packet which is from 192.168.100.0/24 and interface eth1
    • > iptables -A INPUT -i eth1 -s 192.168.100.0/24 -j ACCEPT
  •  Insert a logging rule between the last one which drops packet with iptables something like this would do the trick
    • > iptables -I INPUT (next-to-the-last rule number) -j LOG --log-prefix "blocked packets : "


iptables [-AI 鏈名] [-io 網路介面] [-p 協定] \
> [-s 來源IP/網域] [-d 目標IP/網域] -j [ACCEPT|DROP|REJECT|LOG]
選項與參數:

-S:規則列表
-t:指定表格 ( nat / filter ) 不用t 則預設為 filter
-AI 鏈名:針對某的鏈進行規則的 "插入" "累加"
    -A :新增加一條規則,該規則增加在原本規則的最後面。例如原本已經有四條規則,
         使用 -A 就可以加上第五條規則!
    -I :插入一條規則。如果沒有指定此規則的順序,預設是插入變成第一條規則。
         例如原本有四條規則,使用 -I 則該規則變成第一條,而原本四條變成 2~5
    :有 INPUT, OUTPUT, FORWARD 等,此鏈名稱又與 -io 有關,請看底下。

-io 網路介面:設定封包進出的介面規範
    -i :封包所進入的那個網路介面,例如 eth0, lo 等介面。需與 INPUT 鏈配合;
    -o :封包所傳出的那個網路介面,需與 OUTPUT 鏈配合;

-p 協定:設定此規則適用於哪種封包格式
   主要的封包格式有: tcp, udp, icmp all

-s 來源 IP/網域:設定此規則之封包的來源項目,可指定單純的 IP 或包括網域,例如:
   IP  192.168.0.100
   網域:192.168.0.0/24, 192.168.0.0/255.255.255.0 均可。
   若規範為『不許』時,則加上 ! 即可,例如:
   -s ! 192.168.100.0/24 表示不許 192.168.100.0/24 之封包來源;

-d 目標 IP/網域:同 -s ,只不過這裡指的是目標的 IP 或網域。

-j :後面接動作,主要的動作有接受(ACCEPT)、丟棄(DROP)、拒絕(REJECT)及記錄(LOG)

iptables -L -n -v -x

iptables -N TRAFFIC_ACCT
iptables -I FORWARD -j TRAFFIC_ACCT
iptables -D FORWARD -j TRAFFIC_ACCT
iptables -X TRAFFIC_ACCT


iptables -A TRAFFIC_ACCT -p tcp
iptables -A TRAFFIC_ACCT -p udp
iptables -A TRAFFIC_ACCT -p icmp

Tuesday, November 20, 2012

[Mongrel2] How to write a handler for mongrel2 web server

Mongrel2 is an application, language, and network architecture agnostic web server that focuses on web applications. The most powerful functionality is to use Handler to deal with ZeroMQ message.
There are already some articles talking about Handler and how to get started, for instance:
http://www.ioncannon.net/programming/1384/example-mongrel2-handler-in-ruby/
http://brubeck.io/demos.html

I almost studied 2 days to understand the usage of handler and try a simple handler to respond the request, and finally I finished. So, in this article, I will give a concept about Mongrel2 web server and a simple example.


  • From this diagram, you will quickly realize how the mongrel2 web server works with your handler by what kind of zeromq communication type.
  • Second, I will give the handler example in details:
The conf file ==> mongrel2.conf
brubeck_handler = Handler(
    send_spec='tcp://127.0.0.1:9999',
    send_ident='34f9ceee-cd52-4b7f-b197-88bf2f0ec378',
    recv_spec='tcp://127.0.0.1:9998',
    recv_ident='')

media_dir = Dir(
    base='media/',
    index_file='index.html',
    default_ctype='text/plain')

brubeck_host = Host(
    name="localhost",
    routes={
        '/media/': media_dir,
        '/handlers': brubeck_handler})

brubeck_serv = Server(
    uuid="f400bf85-4538-4f7a-8908-67e313d515c2",
    access_log="/log/mongrel2.access.log",
    error_log="/log/mongrel2.error.log",
    chroot="./",
    default_host="localhost",
    name="brubeck test",
    pid_file="/run/mongrel2.pid",
    port=6767,
    hosts = [brubeck_host]
)

settings = { "zeromq.threads": 1 }

servers = [brubeck_serv]


Load the conf file to mongrel2
> m2sh load -config mongrel2.conf -db the.db

Run the mongrel2 web server
> m2sh start -db the.db -host localhost

Run the simple handler source code ( in Python ) ==> simple_handler.py
> python simple_handler.py

#!/usr/bin/env python
import zmq

ctx = zmq.Context()
pull = ctx.socket(zmq.PULL)
pull.connect("tcp://127.0.0.1:9999")
pub = ctx.socket(zmq.PUB)
pub.connect("tcp://127.0.0.1:9998")

while True:
    msg = pull.recv()
    print msg
    sender_uuid, client_id, request_path, request_message = msg.split(" ", 4)
    ret_content = "Hi...This is Danny..."
    ret_msg = '%s 1:%s, HTTP/1.1 200 OK\r\nContent-Length: %d\r\n\r\n%s' %
    (sender_uuid, client_id , ret_content.__len__(), ret_content)
    pub.send(ret_msg, 0)




Use curl to test it and will get the result as follows:
> curl http://localhost:6767/handlers
Hi...This is Danny...



Monday, November 12, 2012

[ZeroMQ] the study note of the ØMQ guilde -- Part II

Shared Queue ( DEALER and ROUTER sockets)
The only constraint is that services must be stateless, all state being in the request or in some shared storage such as a database.
It then uses zmq-poll[3] to monitor these two sockets for activity and when it has some, it shuttles messages between its two sockets. It doesn't actually manage any queues explicitly — ØMQ does that automatically on each socket.

Built-in Proxy Function
Please see the examle of msgqueue.c that replaces rrbroker.c with built-in proxy function.

Transport Bridging
fig18.png

Handling Errors
In most of the C examples we've seen so far there's been no error handling. Real code should do error handling on every single ØMQ call.
void *context = zmq-ctx-new ();
assert (context);
void *socket = zmq-socket (context, ZMQ-REP);
assert (socket);
int rc = zmq-bind (socket, "tcp://*:5555");
if (rc != 0) {
printf ("E: bind failed: %s\n", strerror (errno));
return -1;
}

We'll use a publish-subscribe model to send kill messages to the workers:
  • The sink creates a PUB socket on a new endpoint.
  • Workers bind their input socket to this endpoint.
  • When the sink detects the end of the batch it sends a kill to its PUB socket.
  • When a worker detects this kill message, it exits.
It doesn't take much new code in the sink:
void *control = zmq-socket (context, ZMQ-PUB);
zmq-bind (control, "tcp://*:5559");

//Send kill signal to workers
zmq-msg-init-data (&message, "KILL", 5);
zmq-msg-send (control, &message, 0);
zmq-msg-close (&message);
Handling Interrupt Signals
s-catch-signals ();
client = zmq-socket (...);
while (!s-interrupted) {
    char *message = s-recv (client);
    if (!message)
        break;          //  Ctrl-C used
}
zmq-close (client);

Detecting Memory Leaks
valgrind: sudo apt-get install valgrind

Multithreading with ØMQ
If you've spent years learning tricks to make your MT code work at all, let alone rapidly, with locks and semaphores and critical sections, you will be disgusted when you realize it was all for nothing. If there's one lesson we've learned from 30+ years of concurrent programming it is: just don't share state.
You should follow some rules to write happy multithreaded code with ØMQ:
  • You must not access the same data from multiple threads. Using classic MT techniques like mutexes are an anti-pattern in ØMQ applications. The only exception to this is a ØMQ context object, which is threadsafe.
  • You must create a ØMQ context for your process, and pass that to all threads that you want to connect via inproc sockets.
  • You may treat threads as separate tasks, with their own context, but these threads cannot communicate over inproc. However they will be easier to break into standalone processes afterwards.
  • You must not share ØMQ sockets between threads. ØMQ sockets are not threadsafe. Technically it's possible to do this, but it demands semaphores, locks, or mutexes. This will make your application slow and fragile. The only place where it's remotely sane to share sockets between threads are in language bindings that need to do magic like garbage collection on sockets.

Singaling between Threads ( PAIR sockets )
the only mechanism that you should use are ØMQ messages.
This is a classic pattern for multithreading with ØMQ ( The Relay Race ) :
  1. Two threads communicate over inproc, using a shared context.
  2. The parent thread creates one socket, binds it to an inproc:// endpoint, and then starts the child thread, passing the context to it.
  3. The child thread creates the second socket, connects it to that inproc:// endpoint, and then signals to the parent thread that it's ready.
fig21.png
mtrelay.c : multithreaded relay in C
For these reasons, PAIR makes the best choice for coordination between pairs of threads.

Node Coordination
syncpub.c: Synchronized publisher in C
  • The publisher knows in advance how many subscribers it expects. This is just a magic number it gets from somewhere.
  • The publisher starts up and waits for all subscribers to connect. This is the node coordination part. Each subscriber subscribes and then tells the publisher it's ready via another socket.
  • When the publisher has all subscribers connected, it starts to publish data.

Zero Copy
ØMQ's message API lets you can send and receive messages directly from and to application buffers without copying data. We call "zero-copy", and it can improve performance in some applications.
There is no way to do zero-copy on receive: ØMQ delivers you a buffer that you can store as long as you wish but it will not write data directly into application buffers.

Pub-Sub Message Envelops
Pub-Sub Envelope with Separate Key
fig23.png
psenvpub.c: Pub-Sub envelop publisher in C
s_sendmore (publisher, "B");
s_send (publisher, "We would like to see this");
psenvsub.c: Pub-Sub envelope subscriber in C
//Read envelope with address
char *address = s_recv (subscriber);
//Read message contents
char *contents = s_recv (subscriber);
printf ("[%s] %s\n", address, contents);

High Water Marks
When you can send messages rapidly from process to process, you soon discover that memory is a precious resource, and one that can be trivially filled up.
The answer for messaging is to set limits on the size of buffers, and then when we reach those limits, take some sensible action. In most cases (not for a subway system, though), the answer is to throw away messages. In a few others, it's to wait.
In ØMQ/2.x the HWM was infinite by default. In ØMQ/3.x it's set to 1,000 by default, which is more sensible.
Some sockets (PUB, PUSH) only have transmit buffers. Some (SUB, PULL, REQ, REP) only have receive buffers. Some (DEALER, ROUTER, PAIR) have both transmit and receive buffers.