Linux Tune Network Stack (Buffers Size) To Increase Networking Performance


, ,


What will happen if clients of TCP/IP stream socket service send data faster than the other client receive and work with it. That is some sort of service has to receive incoming stream and store it until  that service call ‘recv’.


TCP provides flow control . The TCP stack (both on the sender and receiver side) will be able to buffer some data, and this is usually done in the OS kernel.

When the receiver buffers fill up, the sender will know about it, and stop sending more data, eventually leading to the sending application blocking(or otherwise not being able to send more data) until space becomes available again.

Shortly described, every TCP packet(segment) sent includes the size of data that can be buffered – the window size. This means the other end at all times know how much data it can send without the receiver throwing it away because the buffers are full. If the window size becomes 0, buffers are full and no more data will be sent (and in case of the sender being blocking, a send() call will block), Theres procedures for probing whether the tcp window is still 0, so sending can resume again when the data has been consumed.

The default maximum Linux TCP buffer sizes are way too small. TCP memory is calculated automatically based on system memory; you can find the actual values by typing the following commands:
$ cat /proc/sys/net/ipv4/tcp_mem

The default and maximum amount for the receive socket memory:
$ cat /proc/sys/net/core/rmem_default
$ cat /proc/sys/net/core/rmem_max

The default and maximum amount for the send socket memory:
$ cat /proc/sys/net/core/wmem_default
$ cat /proc/sys/net/core/wmem_max
Tune values

Set the max OS send buffer size (wmem) and receive buffer size (rmem) to 12 MB for queues on all protocols. In other words set the amount of memory that is allocated for each TCP socket when it is opened or created while transferring files:
[Warning examples may crash your computer] WARNING! The default value of rmem_max and wmem_max is about 128 KB in most Linux distributions, which may be enough for a low-latency general purpose network environment or for apps such as DNS / Web server. However, if the latency is large, the default size might be too small. Please note that the following settings going to increase memory usage on your server.

# echo ‘net.core.wmem_max=12582912’ >> /etc/sysctl.conf
# echo ‘net.core.rmem_max=12582912’ >> /etc/sysctl.conf

You also need to set minimum size, initial size, and maximum size in bytes:
# echo ‘net.ipv4.tcp_rmem= 10240 87380 12582912’ >> /etc/sysctl.conf
# echo ‘net.ipv4.tcp_wmem= 10240 87380 12582912’ >> /etc/sysctl.conf

Now reload the changes:
# sysctl -p

How to deal with MySQL deadlocks


, ,

In my application, I found Query failed several times with this error report.

‘Deadlock found when trying to get lock; try restarting transaction’

I will not discuss about deadlock and why deadlock occurs. We all know that more than one operations in a single resource make deadlock situation.

So, how to fix this?

At first we have to know the details of the deadlock. I know these two ways.

  1. The latest deadlock can be reviewed using SHOW ENGINE INNODB STATUS command.
  2. With MySQL 5.6, you can enable a new variable innodb_print_all_deadlocks to have all deadlocks in InnoDB recorded in mysqld error log.

You will see like this.

160428 12:13:23
TRANSACTION 766ABFF1, ACTIVE 1 sec starting index read
mysql tables in use 1, locked 1
LOCK WAIT 3 lock struct(s), heap size 1248, 2 row lock(s)
MySQL thread id 3, OS thread handle 0x2ab5dc319940, query id 14760240 root Searching rows for update
UPDATE smsoutbox SET InstanceID = 28 WHERE schedule < NOW() AND ServiceID = '21270_Game_Factory_Daily' AND srcMN = '21270' AND msgStatus = 'QUE' AND InstanceID IS NULL Order by msgID LIMIT 50
RECORD LOCKS space id 0 page no 1662 n bits 168 index `PRIMARY` of table `smsgw_4_0`.`smsoutbox` trx id 766ABFF1 lock_mode X locks rec but not gap waiting
Record lock, heap no 87 PHYSICAL RECORD: n_fields 29; compact format; info bits 0
 0: len 4; hex 00034b45; asc   KE;;
 1: len 6; hex 000043ae5f5e; asc   C _^;;
 2: len 7; hex 430000110f1de6; asc C      ;;
 3: len 5; hex 3231323730; asc 21270;;
 4: len 13; hex 38383031393430343939393437; asc 8801940499947;

TRANSACTION 766ABFC8, ACTIVE 3 sec fetching rows
mysql tables in use 1, locked 1
27209 lock struct(s), heap size 2652600, 1843804 row lock(s)
MySQL thread id 1, OS thread handle 0x2ab5dc081940, query id 14720720 root Updating
UPDATE smsoutbox SET msgStatus = 'que', retryCount = retryCount+1, IsChargingEnable = 0 WHERE msgStatus = 'failed' AND retryCount < 3  and (NOW() - sentTime)> 10 AND InstanceID = 28 AND (CASE WHEN IsChargingEnable = 1 THEN ChargeResponse WHEN IsChargingEnable = 0 THEN '0' END) = '0'
*** (2) HOLDS THE LOCK(S):
RECORD LOCKS space id 0 page no 1662 n bits 168 index `PRIMARY` of table `smsgw_4_0`.`smsoutbox` trx id 766ABFC8 lock_mode X
Record lock, heap no 1 PHYSICAL RECORD: n_fields 1; compact format; info bits 0
 0: len 8; hex 73757072656d756d; asc supremum;;

Record lock, heap no 2 PHYSICAL RECORD: n_fields 29; compact format; info bits 0
 0: len 4; hex 00034af0; asc   J ;;
 1: len 6; hex 000043ae5f5e; asc   C _^;;
 2: len 7; hex 430000110f1445; asc C     E;;
 3: len 4; hex 32353830; asc 2580;;

RECORD LOCKS space id 0 page no 38404 n bits 120 index `PRIMARY` of table `smsgw_4_0`.`smsoutbox` trx id 766ABFC8 lock_mode X waiting
Record lock, heap no 52 PHYSICAL RECORD: n_fields 29; compact format; info bits 0
 0: len 4; hex 0022aa2a; asc  " *;;
 1: len 6; hex 0000766abff2; asc   vj  ;;
 2: len 7; hex c20000019e0110; asc        ;;
 3: len 4; hex 32303038; asc 2008;;


Here you can find those two queries which have made the deadlock with details ( 2nd transactions holds the lock, 1st transaction waits).

Now where is the problem?

Two queries are different, where conditions are also different. So, two queries should not make deadlock. But it creates deadlock. Now after some studies I found the main reason.

InnoDB does not remember the exact WHERE condition, but only knows which index ranges were scanned. The locks are normally next-key locks that also block inserts into the gap immediately before the record. (

If you have no indexes suitable for your statement and MySQL must scan the entire table to process the statement, every row of the table becomes locked, which in turn blocks all inserts by other users to the table. It is important to create good indexes so that your queries do not unnecessarily scan many rows.

As both of my queries use same index (index ‘Primary’), deadlock is created. So, indexing is the solution. I have added index.

If two queries use different index, they will not make deadlock.

Now how to be sure that two queries are using different index? Use EXPLAIN keyword in front of your queries and see which key is using.

WHERE SCHEDULE < NOW() AND ServiceID = ‘21270_Game_Factory_Daily’ AND srcMN = ‘21270’ AND msgStatus = ‘QUE’ AND InstanceID IS NULL ORDER BY msgID LIMIT 50.Capture

So, if you are working with InnoDB, you have to think about index to avoid deadlock.

Thank you.

You can find some details from this link also.

Unicode in MySQL and C



For my application, I have to save Unicode message in MySQL database. And my C application will pick that value.

In MySQL part,

# For each database:
ALTER DATABASE database_name CHARACTER SET = utf8mb4 COLLATE = utf8mb4_unicode_ci;
# For each table:
ALTER TABLE table_name CONVERT TO CHARACTER SET utf8mb4 COLLATE utf8mb4_unicode_ci;
# For each column:
ALTER TABLE table_name CHANGE column_name column_name VARCHAR(191) CHARACTER SET utf8mb4 COLLATE utf8mb4_unicode_ci;

# (Don’t blindly copy-paste this! You have to use your own values.)

Now check.

SHOW VARIABLES WHERE Variable_name LIKE 'character\_set\_%' OR Variable_name LIKE 'collation%';

For MySQL part, You can get details from


Now, what I have to do in my C application

 // these two lines extra for unicode
 mysql_options(&mysql, MYSQL_SET_CHARSET_NAME, "utf8"); 
 mysql_options(&mysql, MYSQL_INIT_COMMAND, "SET NAMES utf8"); 


Now you can use normal procedure.

ret = mysql_query(con, your_query);

Binary to Decimal and Decimal to Binary Conversion


, , ,

#include <stdio.h>
#include <math.h>

int binary_to_decimal(int binary) /* Function to convert binary to decimal.*/
int decimal = 0, i = 0, reminder;
while (binary != 0)
reminder = binary % 10;
binary /= 10;
decimal += reminder*pow(2.0,i);
return decimal;

int decimal_to_binary(int decimal) /* Function to convert decimal to binary.*/
int reminder, i = 1, binary = 0;
while (decimal != 0)
reminder = decimal%2;
decimal /= 2;
binary += reminder*i;
i *= 10;
return binary;

Introduction to Android Native Development


, , ,

When I had started Android Native Development, I had to face some problems. So, I am writing this tutorial for the beginners who is just starting.

Environment Setup:
1. Eclipse 4.2
2. JDk7
3. Android SDK
4. CDT Plugin
5. ADT Plugin
6. NDK

Now Install JDK, Eclipse. Configure SDK location. Install CDT Plugin. Then Install ADT Plugin (If you don’t install CDT, you won’t install NDK plugin).
After you install NDK plugin, you will find a new sub-menu option of Android(Windows-> Preference->Android). Set NDK location (be sure the path can’t contain a space).

Note: In some tutorial you may find that, they mentioned about Cygwin or others to compile NDK. But latest, even I used android-ndk-r8, I didn’t need those.

Create Project
1. Create a new Android project.
2. Right click on project explorer. you find “Android tools” in the menu.
3. Select “Add Native Support…”. Write your library name. You will find a new folder in project explorer (jni). jni contains a cpp file and
4. Now write your method in cpp.
JNIEXPORT return_type JNICALL Java_yourPackage_YourJavaClass_MethodName(JNIEnv *env, jobject thiz)

JNIEnv *env, jobject thiz MUST be in parameters. If you need more, you can add after those. As it is cpp file, you need to use extern “C”.

Here I used.
extern “C” jint Java_com_example_ndktest_MainActivity_AddValue(JNIEnv *env,
jobject thiz,
jint nValue1,
jint nValue2)

In your MainActivity class, You should write.
System.loadLibrary(“MathSDK”); // is the lib name

public native int AddValue(int nValue1, int nValue2);

Now You can use AddValue Method as a member method of MainActivity class.
You don’t need to write anything till now in

Now build and run.

Convert char[] to hex string

char strInput[] = "yourchardata";
char chHex[3] = "";
int nLength = strlen(strInput);

char* chResut = new char[(nLength*2) + 1];
memset(chResut, 0, (nLength*2) + 1);

for (int i = 0; i < nLength; i++)
sprintf(chHex, "%02X", strInput[i] & 0x00FF);
memcpy(&(chResut[i*2]), chHex, 2);

printf("\n%s", chResut);
delete chResut;
chResut = NULL;

Create and Delete Thread in MFC



CWinThread* m_pThread;

// Creating Thread
if (NULL == m_pThread)
m_pThread = AfxBeginThread(ThreadProc,

m_pThread->m_bAutoDelete = false;

TerminateThread forces another thread to exit. You should avoid calling it at all costs as it will stop a thread dead in it’s tracks without any chance to cleanup. This includes any CRT memory allocated.

ExitThread is for the currently running thread to stop itself nice and cleanly. When you called it above, you likely forced the main (UI) thread to exit and likely left the running threads still lingering around. Hence, your program was still running as evidenced in Task Manager. GetExitCodeThread was also likely failing since the threads had not actually exited.

But the right way to stop a thread is to cleanly signal by any clean means necessary that it should exit. Then allow the threads to exit on their own before allowing the main thread to exit.

// Deleting thread
WaitForSingleObject(m_pThread->m_hThread, INFINITE);

// get the thread's exit code (I'm not sure why you need it)
DWORD dwExitCode;
BOOL bRet = GetExitCodeThread(m_pThread->m_hThread, &dwExitCode);

// cleanup the thread
m_pThread->m_hThread = NULL;

delete m_pThread;
m_pThread = NULL;

Get console output from exe in C++


, ,

I have an executable file which prints “Hello World”;  File name “PrintOutput.exe”.

I have to get the console output that is “Hello World” in C++ project. How will I do that?

int _tmain(int argc, _TCHAR* argv[])
char   psBuffer[128];
FILE   *pPipe;

pPipe = _popen( “PrintOutput.exe”, “rt” );
if (NULL == pPipe)
printf(“error input”);
return 0;

while(NULL != fgets(psBuffer, 128, pPipe))
printf(psBuffer); // print the console output of “PrintOutput.exe”


_pclose( pPipe );



Get console output from exe in php


, ,

I have an executable file which prints “Hello World”;  File name “PrintOutput.exe”.

I have to get the console output that is “Hello World” in my PHP page. How will I do that?


// put your code here

$output = shell_exec(“PrintOutput.exe”);

echo $output;     // it will print “Hello World”




Unsorted map


, ,

In C++, most of the programmers use map to store elements formed by a combination of a key value and a mapped value, following a specific order.

But for some cases, I needed an unsorted map. I am explaining.

I have some data,of string type such as date, account number, account name, balance, debit amount, credit amount etc. All are key values here. I have to store values associated with these.

Problem is, map is sorting my data. that is account name, account number, balance, date order. I am not getting data in my inserted order.

So, How did I solve this problem?

I used.

struct Cell
string keyValue;
string data;

vector<Cell> cells;

One of the most benefits is, vector is faster than map.