MySQL HandlerSocket under Ubuntu

Starting with the great work of Yoshinori-san Using MySQL as a NoSQL – A story for exceeding 750,000 qps on a commodity server and Golan Zakai who posted Installing Dena’s HandlerSocket NoSQL plugin for MySQL on Centos I configured and tested HandlerSocket under Ubuntu 10.04 64bit.

NOTE: This machine already compiles MySQL and Drizzle. You should refer to appropriate source compile instructions for necessary dependencies.

# Get Software
cd /some/path
export DIR=`pwd`
wget http://download.github.com/ahiguti-HandlerSocket-Plugin-for-MySQL-1.0.6-10-gd032ec0.tar.gz
wget http://mysql.mirror.iweb.ca/Downloads/MySQL-5.1/mysql-5.1.52.tar.gz
wget  http://mysql.mirror.iweb.ca/Downloads/MySQL-5.1/mysql-5.1.52-linux-x86_64-glibc23.tar.gz

# Install MySQL Binaries
tar xvfz mysql-5.1.52-linux-x86_64-glibc23.tar.gz
ln -s mysql-5.1.52-linux-x86_64-glibc23 mysql5152
rm -rf mysql5152/mysql-test
rm -rf mysql5152/sql-bench
cd mysql5152
scripts/mysql_install_db
bin/mysqld_safe &

# Install MySQL Source
tar xvfz mysql-5.1.52.tar.gz

# Install HandlerSocket
tar xvfz ahiguti-HandlerSocket-Plugin-for-MySQL-1.0.6-10-gd032ec0.tar.gz
cd ahiguti-HandlerSocket-Plugin-for-MySQL-d032ec0/
./autogen.sh
./configure --with-mysql-source=${DIR}/mysql-5.1.51 --with-mysql-bindir=${DIR}/mysql5152/bin
make
sudo make install

# libtool: install: /usr/bin/install -c .libs/handlersocket.lai /home/rbradfor/projects/handlersocket/mysql5152/lib/plugin/handlersocket.la

# Install the Perl dependency
cd perl-Net-HandlerSocket
perl Makefile.PL
Checking if your kit is complete...
Warning: the following files are missing in your kit:
	lib/HandlerSocket.pm
Please inform the author.
Writing Makefile for Net::HandlerSocket
# For some reason I had to run this twice, after error I check, saw nothing wrong, ran a second time and it worked ???
perl Makefile.PL
Writing Makefile for Net::HandlerSocket
make
sudo make install

Files found in blib/arch: installing files in blib/lib into architecture dependent library tree
Installing /usr/local/lib/perl/5.10.1/auto/Net/HandlerSocket/HandlerSocket.so
Installing /usr/local/lib/perl/5.10.1/auto/Net/HandlerSocket/HandlerSocket.bs
Installing /usr/local/lib/perl/5.10.1/Net/HandlerSocket.pm
Installing /usr/local/lib/perl/5.10.1/Net/HandlerSocket/Pool.pm
Installing /usr/local/man/man3/Net::HandlerSocket.3pm
Appending installation info to /usr/local/lib/perl/5.10.1/perllocal.pod

cp lib/Net/HandlerSocket.pm blib/lib/Net/HandlerSocket.pm
cp lib/Net/HandlerSocket/Pool.pm blib/lib/Net/HandlerSocket/Pool.pm
/usr/bin/perl /usr/share/perl/5.10/ExtUtils/xsubpp  -typemap /usr/share/perl/5.10/ExtUtils/typemap  HandlerSocket.xs > HandlerSocket.xsc && mv HandlerSocket.xsc HandlerSocket.c
Please specify prototyping behavior for HandlerSocket.xs (see perlxs manual)
g++ -fPIC -c  -I. -I../libhsclient -D_REENTRANT -D_GNU_SOURCE -DDEBIAN -fno-strict-aliasing -pipe -fstack-protector -I/usr/local/include -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -g -O3 -Wall -Wno-unused   -DVERSION="0.01" -DXS_VERSION="0.01" -fPIC "-I/usr/lib/perl/5.10/CORE"   HandlerSocket.c
Running Mkbootstrap for Net::HandlerSocket ()
chmod 644 HandlerSocket.bs
rm -f blib/arch/auto/Net/HandlerSocket/HandlerSocket.so
LD_RUN_PATH="/usr/local/lib" g++ -fPIC  -shared -O2 -g -L/usr/local/lib -fstack-protector HandlerSocket.o  -o blib/arch/auto/Net/HandlerSocket/HandlerSocket.so 	
	   -lhsclient  	

chmod 755 blib/arch/auto/Net/HandlerSocket/HandlerSocket.so
cp HandlerSocket.bs blib/arch/auto/Net/HandlerSocket/HandlerSocket.bs
chmod 644 blib/arch/auto/Net/HandlerSocket/HandlerSocket.bs
Manifying blib/man3/Net::HandlerSocket.3pm

Now all built, it’s as simple as the following to install:

${DIR}/mysql5152/bin/mysql -uroot
mysql> INSTALL PLUGIN HandlerSocket SONAME 'handlersocket.so';
mysql> SHOW PLUGINS;
+---------------+----------+----------------+------------------+---------+
| Name          | Status   | Type           | Library          | License |
+---------------+----------+----------------+------------------+---------+
| binlog        | ACTIVE   | STORAGE ENGINE | NULL             | GPL     |
| partition     | ACTIVE   | STORAGE ENGINE | NULL             | GPL     |
| ARCHIVE       | ACTIVE   | STORAGE ENGINE | NULL             | GPL     |
| BLACKHOLE     | ACTIVE   | STORAGE ENGINE | NULL             | GPL     |
| CSV           | ACTIVE   | STORAGE ENGINE | NULL             | GPL     |
| FEDERATED     | DISABLED | STORAGE ENGINE | NULL             | GPL     |
| MEMORY        | ACTIVE   | STORAGE ENGINE | NULL             | GPL     |
| InnoDB        | ACTIVE   | STORAGE ENGINE | NULL             | GPL     |
| MyISAM        | ACTIVE   | STORAGE ENGINE | NULL             | GPL     |
| MRG_MYISAM    | ACTIVE   | STORAGE ENGINE | NULL             | GPL     |
| handlersocket | ACTIVE   | DAEMON         | handlersocket.so | BSD     |
+---------------+----------+----------------+------------------+---------+
11 rows in set (0.00 sec)

However, I found it not actually running even after restarting MySQL. You need the following configuration

cd ${DIR}/mysql5152
echo "[mysqld]
plugin-load=handlersocket.so
loose_handlersocket_port = 9998 # the port number to bind to (for read requests)
loose_handlersocket_port_wr = 9999 # the port number to bind to (for write requests)
loose_handlersocket_threads = 16 # the number of worker threads (for read requests)
loose_handlersocket_threads_wr = 1 # the number of worker threads (for write requests)" >> my.cnf

After restarting you get the following confirmation.

mysql>SHOW PROCESSLIST;
+----+-------------+-----------------+---------------+---------+------+-------------------------------------------+------------------+
| Id | User        | Host            | db            | Command | Time | State                                     | Info             |
+----+-------------+-----------------+---------------+---------+------+-------------------------------------------+------------------+
|  1 | system user | connecting host | NULL          | Connect | NULL | handlersocket: mode=rd, 0 conns, 0 active | NULL             |
|  2 | system user | connecting host | NULL          | Connect | NULL | handlersocket: mode=rd, 0 conns, 0 active | NULL             |
|  3 | system user | connecting host | NULL          | Connect | NULL | handlersocket: mode=rd, 0 conns, 0 active | NULL             |
|  4 | system user | connecting host | NULL          | Connect | NULL | handlersocket: mode=rd, 0 conns, 0 active | NULL             |
|  5 | system user | connecting host | NULL          | Connect | NULL | handlersocket: mode=rd, 0 conns, 0 active | NULL             |
|  6 | system user | connecting host | NULL          | Connect | NULL | handlersocket: mode=rd, 0 conns, 0 active | NULL             |
|  7 | system user | connecting host | NULL          | Connect | NULL | handlersocket: mode=rd, 0 conns, 0 active | NULL             |
|  8 | system user | connecting host | NULL          | Connect | NULL | handlersocket: mode=rd, 0 conns, 0 active | NULL             |
|  9 | system user | connecting host | NULL          | Connect | NULL | handlersocket: mode=rd, 0 conns, 0 active | NULL             |
| 10 | system user | connecting host | NULL          | Connect | NULL | handlersocket: mode=rd, 0 conns, 0 active | NULL             |
| 11 | system user | connecting host | NULL          | Connect | NULL | handlersocket: mode=rd, 0 conns, 0 active | NULL             |
| 12 | system user | connecting host | NULL          | Connect | NULL | handlersocket: mode=rd, 0 conns, 0 active | NULL             |
| 13 | system user | connecting host | NULL          | Connect | NULL | handlersocket: mode=rd, 0 conns, 0 active | NULL             |
| 14 | system user | connecting host | NULL          | Connect | NULL | handlersocket: mode=rd, 0 conns, 0 active | NULL             |
| 15 | system user | connecting host | NULL          | Connect | NULL | handlersocket: mode=rd, 0 conns, 0 active | NULL             |
| 16 | system user | connecting host | NULL          | Connect | NULL | handlersocket: mode=rd, 0 conns, 0 active | NULL             |
| 17 | system user | connecting host | handlersocket | Connect | NULL | handlersocket: mode=wr, 0 conns, 0 active | NULL             |
| 18 | root        | localhost       | NULL          | Query   |    0 | NULL                                      | SHOW PROCESSLIST |
+----+-------------+-----------------+---------------+---------+------+-------------------------------------------+------------------+
18 rows in set (0.00 sec)

My Test Data

USE test
CREATE TABLE user (
  user_id INT UNSIGNED NOT NULL,
  name VARCHAR(50) NOT NULL,
  email VARCHAR(255) NOT NULL,
  created TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP,
  PRIMARY KEY(user_id)
) ENGINE=InnoDB;

INSERT INTO user (user_id,name, email) VALUES
(100,'Mickey Mouse','[email protected]'),
(101,'Marvin the Martian','[email protected]'),
(102,'Spiderman','[email protected]');

SELECT user_id,name,email FROM user WHERE user_id=101;
+---------+--------------------+---------------------+
| user_id | name               | email               |
+---------+--------------------+---------------------+
|     101 | Marvin the Martian | [email protected] |
+---------+--------------------+---------------------+

My test retrieval program is basically unchanged from the original post.

$ cat retrieve.pl
#!/usr/bin/perl

use strict;
use warnings;
use Net::HandlerSocket;

#1. establishing a connection
my $args = { host => '10.0.0.6', port => 9998 };
my $hs = new Net::HandlerSocket($args);

#2. initializing an index so that we can use in main logics.
 # MySQL tables will be opened here (if not opened)
my $res = $hs->open_index(0, 'test', 'user', 'PRIMARY',
    'name,email,created');
die $hs->get_error() if $res != 0;

#3. main logic
 #fetching rows by id
 #execute_single (index id, cond, cond value, max rows, offset)
$res = $hs->execute_single(0, '=', [ '101' ], 1, 0);
die $hs->get_error() if $res->[0] != 0;
shift(@$res);
for (my $row = 0; $row < 1; ++$row) {
  my $user_name= $res->[$row + 0];
  my $user_email= $res->[$row + 1];
  my $created= $res->[$row + 2];
  print "$user_namet$user_emailt$createdn";
}

#4. closing the connection
$hs->close();

If your HandlerSocket is not running you will see:

$ perl retrieve.pl
FATAL_EXIT: getaddrinfo failed: ip_to_remote_host:9998
$ telnet 10.0.0.6 9998
Trying 10.0.0.6...
telnet: Unable to connect to remote host: Connection refused

When working

$ telnet 10.0.0.6 9998
Trying 10.0.0.6...
Connected to 10.0.0.6.
Escape character is '^]'.
P	0	test	user	PRIMARY	name,email,created
0	1
0	=	1	101
0	3	Marvin the Martian	[email protected]	2010-11-05 13:11:34


$ perl retrieve.pl
Marvin the Martian	[email protected]	2010-11-05 13:11:34

Again, great work by Yoshinori-san to have this code as production deployable. Now to run some real benchmarks.

Compiling libdrizzle

Compiling libdrizzle is a rather trivial task. The following are the steps I undertook on Ubuntu 8.10 Intrepid 32 bit.

There was one pre-requisite from the most basic installed developer tools.

sudo apt-get install -y  automake
bzr clone lp:libdrizzle
cd libdrizzle
./config/autorun.sh
./configure
make
make install

And there they are:

$ ls -l /usr/local/lib/libdrizzle*
-rw-r--r-- 1 root root 1122710 2009-04-26 18:10 /usr/local/lib/libdrizzle.a
-rwxr-xr-x 1 root root     940 2009-04-26 18:10 /usr/local/lib/libdrizzle.la
lrwxrwxrwx 1 root root      19 2009-04-26 18:10 /usr/local/lib/libdrizzle.so -> libdrizzle.so.0.0.2
lrwxrwxrwx 1 root root      19 2009-04-26 18:10 /usr/local/lib/libdrizzle.so.0 -> libdrizzle.so.0.0.2
-rwxr-xr-x 1 root root 1003734 2009-04-26 18:10 /usr/local/lib/libdrizzle.so.0.0.2

I added the following to enable other programs using libdrizzle to find the libraries in the system path.

echo "/usr/local/lib" > /etc/ld.so.conf.d/drizzle.conf
ldconfig

Drizzle/bzr dependency

A number of developers had problems on Friday at the Drizzle Developer Day with compiling bzr. The distro in question I was helping with was CentOS 5 32-bit. I had no issues on CentOS 5 64bit.

Today while creating the first deployed Drizzle AWS AMI I discovered the same problem using Ubuntu 8.10 Intrepid 32 bit.

The solution was actually rather trivial. Installing the python-dev package solved the problem.

apt-get install python-dev
Bzr 1.13.1 Compiling error

building 'bzrlib._btree_serializer_c' extension
gcc -pthread -fno-strict-aliasing -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -fPIC -I/usr/include/python2.5 -c bzrlib/_btree_serializer_c.c -o build/temp.linux-i686-2.5/bzrlib/_btree_serializer_c.o
bzrlib/_btree_serializer_c.c:4:20: error: Python.h: No such file or directory
bzrlib/_btree_serializer_c.c:5:26: error: structmember.h: No such file or directory
bzrlib/_btree_serializer_c.c:35: error: expected specifier-qualifier-list before ‘PyObject’
....
bzrlib/_btree_serializer_c.c:1651: error: request for member ‘f_lineno’ in something not a structure or union
bzrlib/_btree_serializer_c.c:1651: warning: statement with no effect
bzrlib/_btree_serializer_c.c:1652: warning: implicit declaration of function ‘PyTraceBack_Here’

  Cannot build extension "bzrlib._btree_serializer_c".
  Use "build_ext --allow-python-fallback" to use slower python implementations instead.

error: command 'gcc' failed with exit status 1

A Drizzle update – Running version 2009.03.970-development

I’ve not looked at compiling and running Drizzle on my server for the past four weeks. Well overdue time for a check and see how it’s going. I saw in today’s planet.mysql.com by Eric Day a new dependency is needed. libdrizzle 0.2.0 now in Drizzle is now required, so I started there.

cd ~/bzr
bzr branch lp:libdrizzle
cd libdrizzle
./config/autorun.sh
./configure
make
sudo make install

No problems there, also documented at the Drizzle Wiki. Great to see the docs up to date. I see my old work on starting the compiling page still relevant. Tested on CentOS 5 and Mac OS/X 10.5

Compiling drizzle was not much more difficult.

cd ~/bzr/drizzle
bzr update
make distclean
./config/autorun.sh
./configure --prefix=/home/drizzle/deploy
make
make install

The problems happened when I started drizzle. Initially I was using bin/drizzled_safe, but it was recommended via IRC#drizzle I stick with sbin/drizzled

sbin/drizzled &
error while loading shared libraries: libprotobuf.so.2: cannot open shared object file: No such file or directory

An investigation of Google Proto Buffers.

$ protoc --version
libprotoc 2.0.2

I see that protobuf 2.0.3 is now available, but this was not the problem.

I got around the problem by specifying the current library path:

$ LD_LIBRARY_PATH=/usr/local/lib sbin/drizzled &

I corrected this problem by adding /usr/local/lib to the default ld path, both the libdrizzle and libprotobuf libs are located there.

$ echo "/usr/local/lib" > /etc/ld.so.conf.d/drizzle.conf
$ ldconfig
$ ls -l /usr/local/lib
total 37240
-rw-r--r-- 1 root root  1194602 Mar 31 17:42 libdrizzle.a
-rwxr-xr-x 1 root root      940 Mar 31 17:42 libdrizzle.la
lrwxrwxrwx 1 root root       19 Mar 31 17:42 libdrizzle.so -> libdrizzle.so.0.0.2
lrwxrwxrwx 1 root root       19 Mar 31 17:42 libdrizzle.so.0 -> libdrizzle.so.0.0.2
-rwxr-xr-x 1 root root  1117979 Mar 31 17:42 libdrizzle.so.0.0.2
-rw-r--r-- 1 root root 12199302 Nov 30 23:32 libprotobuf.a
-rwxr-xr-x 1 root root      836 Nov 30 23:32 libprotobuf.la
lrwxrwxrwx 1 root root       20 Nov 30 23:32 libprotobuf.so -> libprotobuf.so.2.0.0
lrwxrwxrwx 1 root root       20 Aug 27  2008 libprotobuf.so.0 -> libprotobuf.so.0.0.0
-rwxr-xr-x 1 root root  5027949 Aug 27  2008 libprotobuf.so.0.0.0
lrwxrwxrwx 1 root root       20 Nov 30 23:32 libprotobuf.so.2 -> libprotobuf.so.2.0.0
-rwxr-xr-x 1 root root  5586965 Nov 30 23:32 libprotobuf.so.2.0.0
-rw-r--r-- 1 root root  9264068 Nov 30 23:32 libprotoc.a
-rwxr-xr-x 1 root root      852 Nov 30 23:32 libprotoc.la
lrwxrwxrwx 1 root root       18 Nov 30 23:32 libprotoc.so -> libprotoc.so.0.0.0
lrwxrwxrwx 1 root root       18 Nov 30 23:32 libprotoc.so.0 -> libprotoc.so.0.0.0
-rwxr-xr-x 1 root root  3645396 Nov 30 23:32 libprotoc.so.0.0.0
drwxr-xr-x 2 root root     4096 Mar 31 17:42 pkgconfig

Starting

$ sbin/drizzled &
InnoDB: The InnoDB memory heap is disabled
InnoDB: Mutexes and rw_locks use GCC atomic builtins.
090331 18:38:08  InnoDB: highest supported file format is Barracuda.
InnoDB: The log sequence number in ibdata files does not match
InnoDB: the log sequence number in the ib_logfiles!
090331 18:38:08  InnoDB: Database was not shut down normally!
InnoDB: Starting crash recovery.
InnoDB: Reading tablespace information from the .ibd files...
InnoDB: Restoring possible half-written data pages from the doublewrite
InnoDB: buffer...
090331 18:38:08 InnoDB Plugin 1.0.3 started; log sequence number 46419
sbin/drizzled: ready for connections.
Version: '2009.03.970-development'  socket: ''  port: 4427  Source distribution

Verifying

$ bin/drizzle -uroot
Welcome to the Drizzle client..  Commands end with ; or g.
Your Drizzle connection id is 1
Server version: 2009.03.970-development Source distribution
Type 'help;' or 'h' for help. Type 'c' to clear the buffer.
drizzle> select version();
+-------------------------+
| version()               |
+-------------------------+
| 2009.03.970-development |
+-------------------------+
1 row in set (0.00 sec)
drizzle> exit

Sweet! Now to try some testing & benchmarking before the barrage of conferences next month, 2009 MySQL Camp, Percona Performance Conference and MySQL Conference & Expo.

I’m going to check out The Juice Database Benchmark next as a more realistic benchmark to DBT2 and sysbench.

My Ideas for MySQL Camp III

Diary: January 21st 2008 – Martin Luther King Day (Day doctor’s practices are closed BTW.)

“I have a dream”, poetic . Actually I have thumping 5 day straight headache but that’s another story.

I have a dream for MySQL Camp III. A 48 hour Global Hackfest. I ran this by Jay over Thanksgiving, to get back to more the purpose of the Camp, for hackers, coders and the very experience to get to together to share their skills, and for those at the top of our respective game to learn just a little more. MySQL Camp II was a success to attendees in general, but of little value to the experts.

I hope to get us middle to advanced ground. Here is an overview.

  • 48 hour event
  • Say 12pm Friday GMT to 12pm Sunday GMT
  • Global meeting points of two or more to work together, sleep, eat and play together (in fact the goal is not to attend alone, you should really try to get to at least the closest person that’s also attending, it’s designed to be a distributed group event)
  • 4 key areas
    • Getting Starrted. Getting the code, understanding the basics of compiling – Linux, Mac OS/X and Windoze. CLI & IDE debugging options
    • A&D. Analysis of problems, selected bugs, existing patches etc, reviewing procedures, documentation requirements, important of test plans etc. Laying down a plan on what’s going to happen, how long it’s going to take etc.
    • Doing it. Taking a working developer environment, and a set plan, and executing to completion
    • Reviewing it. Getting an insite into MySQL and how bug fixes, community contributions etc are submitted, reviewed, proposed and received
  • I would anticipate we would run say 2 or 3 tracks of these 4 points, so we would repeat stuff, perhaps a different problem, but this enables you to get a real grip, as well as cater for the 24hr cycle.
  • I could see the Doing it interesting, perhaps depending on attendees either a mentor process where a code guru could instruct some youngerlings, or multiple teams working in parallel on the same problem, a little bit of competition.
  • Code base line version and list of bugs/features to be looked at to be pre-determined, so we have a clear structure during the event. This will be proposal format, and may include for example back porting patches for example.
  • We will definitely be having some prizes and some fun, it’s going to be one of those work to 3am in the morning weekends regardless of where you live.

I expect to run this format of MySQL Camp twice, the first to work out any serious problems. I had hoped in November last year to get this before UC2008, fat chance, but I’m proposing a MySQL UC2008 a BOF on the subject.

I had spoken with good friend Farshan Mashraqi that had done some good Sun webinars recently and he was seeking a contact to see if Sun could donate the time/bandwidth for the electronic component or even sponsor. (This was clearly discussed late last year) Seems now this may be easier, or harder with Sun’s involvement.

I have a lot more details, but I want to get this out there into the world, and get some feedback first.
I’d like people’s feedback. Here are 5 questions to start with.

  1. Do you think it’s a good idea?
  2. Would you attended/participate?
  3. Could you contribute in some why? What?
  4. The one thing that appeals the most on the concept?
  5. The one thing that appeals the least/lacks/needs on the concept?

Brisbane Users Group – MySQL Hackfest

Last night we had a number of keen souls at the Brisbane MySQL User Group. I was very impressed to see the majority of people with laptops at hand.

You can download my slides and code examples at my Articles page.

In our hands on Hackfest tutorial we created the new command SHOW USERGROUPS. Before anybody makes a comment, it was stated in the presentation that this command was made a dummy one, and is a poor candidate for two reasons.

  1. The results should be more dynamic, rather then hardcoded into the source tree
  2. USERGROUPS is not an ideal name due to comparisions to Users, Groups, Roles etc

Still it was productive, here was the outcome of our work for the evening.

mysql> SHOW USERGROUPS;
+----------------------+-------------------------+-----------------------------+
| Name                 | Location                | Website                     |
+----------------------+-------------------------+-----------------------------+
| Brisbane Users Group | Brisbane, QLD Australia | http://mysql.meetup.com/84/ |
| Boston Users Group   | Boston, USA             |                             |
+----------------------+-------------------------+-----------------------------+
2 rows in set (0.00 sec)

The trailing slash ‘/’ Hitcho’s contribution :)

mysql> SHOW AUTHORS;
+--------------------------------+---------------------------------------+----------------------------------------------------------------------+
| Name                           | Location                              | Comment                             |
+--------------------------------+---------------------------------------+----------------------------------------------------------------------+
| Brian (Krow) Aker              | Seattle, WA, USA                      | Architecture, archive, federated, bunch of little stuff :)           |
...
 Brisbane Users Group           | Brisbane, QLD Australia               | New Command SHOW USERGROUPS                             |
+--------------------------------+---------------------------------------+----------------------------------------------------------------------+
75 rows in set (0.05 sec)

I’ll be publishing more beginner tutorials in my MySQL Compiling Category over time.

Update Check out http://www.arabx.com.au/hackfest.diff.htm for a diff of files to produce this output.

Compiling MySQL Tutorial 3 – Debugging Output

Continuing on from Tutorial 2.

When reviewing the 2.1. C/C++ Coding Guidelines for MySQL, you will see that the MySQL Source uses within the C/C++ code DBUG (Fred Fish’s debug library). This is one of the 3rd party open source products that are documented in the Internals 1.4. The Open-source Directories.

You will also find some usage in the MySQL Manual E.3 The DBUG Package. So enough talk, how do you use this.

DBUG

You get the DBUG output by running a mysqld debug version with the argument –debug. The output for the SHOW AUTHORS commands is:

T@9190304: >dispatch_command
T@9190304: | >alloc_root
..
T@9190304: | query: SHOW AUTHORS
T@9190304: | >mysql_parse
T@9190304: | | >mysql_init_query
T@9190304: | | | >lex_start
T@9190304: | | | | >alloc_root
T@9190304: | | | | | enter: root: 0x9590b38
T@9190304: | | | | | exit: ptr: 0x95c1c70
T@9190304: | | | | <alloc_root
T@9190304: | | | <lex_start
T@9190304: | | | >mysql_reset_thd_for_next_command
T@9190304: | | | <mysql_reset_thd_for_next_command
T@9190304: | | <mysql_init_query
T@9190304: | | >Query_cache::send_result_to_client
T@9190304: | | <query_cache ::send_result_to_client
T@9190304: | | >mysql_execute_command
T@9190304: | | | >mysqld_show_authors
T@9190304: | | | | >alloc_root
T@9190304: | | | | | enter: root: 0x9590b38
T@9190304: | | | | | exit: ptr: 0x95c1c78
T@9190304: | | | | <alloc_root
...
T@9190304: | | | | >send_fields
T@9190304: | | | | | packet_header: Memory: 0x8c2480  Bytes: (4) 01 00 00 01
T@9190304: | | | | | >alloc_root
T@9190304: | | | | >Protocol::write
T@9190304: | | | | <protocol ::write
...
T@9190304: | | | | >send_eof
T@9190304: | | | | | packet_header: Memory: 0x8c2550  Bytes: (4) 05 00 00 51
T@9190304: | | | | | >net_flush
T@9190304: | | | | | | >vio_is_blocking
T@9190304: | | | | | | | exit: 0
T@9190304: | | | | | | <vio_is_blocking
T@9190304: | | | | | | >net_real_write
T@9190304: | | | | | | | >vio_write
T@9190304: | | | | | | | | enter: sd: 23, buf: 0x0x95b9bd0, size: 5107
T@9190304: | | | | | | | | exit: 5107
T@9190304: | | | | | | | <vio_write
T@9190304: | | | | | | <net_real_write
T@9190304: | | | | | <net_flush
T@9190304: | | | | | info: EOF sent, so no more error sending allowed
T@9190304: | | | | <send_eof
T@9190304: | | | <mysqld_show_authors
T@9190304: | | <mysql_execute_command
T@9190304: | | >query_cache_end_of_result
T@9190304: | | <query_cache_end_of_result
...
T@9190304: | <mysql_parse
T@9190304: | info: query ready
T@9190304: | >log_slow_statement
T@9190304: | <log_slow_statement
T@9190304: >dispatch_command

Using some of the debug options you get the following output with –debug=d,info,error,query,general,where:O,/tmp/mysqld.trace

create_new_thread: info: creating thread 3
create_new_thread: info: Thread created
?func: info: handle_one_connection called by thread 3
?func: info: New connection received on socket (23)
?func: info: Host: localhost
?func: info: vio_read returned -1,  errno: 11
thr_alarm: info: reschedule
process_alarm: info: sig: 14  active alarms: 1
?func: info: client_character_set: 8
check_user: info: Capabilities: 238213  packet_length: 16777216  Host: 'localhost'  Login user: 'rbradfor' Priv_user: ''  Using password: no Access: 0  db: '*none*'
send_ok: info: affected_rows: 0  id: 0  status: 2  warning_count: 0
send_ok: info: OK sent, so no more error sending allowed
do_command: info: vio_read returned -1,  errno: 11
thr_alarm: info: reschedule
process_alarm: info: sig: 14  active alarms: 1
do_command: info: Command on socket (23) = 3 (Query)
dispatch_command: query: SHOW AUTHORS
send_eof: info: EOF sent, so no more error sending allowed
dispatch_command: info: query ready
do_command: info: vio_read returned -1,  errno: 11
thr_alarm: info: reschedule
process_alarm: info: sig: 14  active alarms: 1
do_command: info: Command on socket (23) = 1 (Quit)
~THD(): info: freeing security context
end_thread: info: sending a broadcast
end_thread: info: unlocked thread_count mutex


I have found the following format good for comparing with the MySQL Source Code. —debug=d:N:F:L:O,/tmp/mysqld.trace

   49:   sql_parse.cc:  1536: do_command: info: Command on socket (23) = 3 (Query)
   50:     my_alloc.c:   172: alloc_root: enter: root: 0x9068b38
   51:     my_alloc.c:   219: alloc_root: exit: ptr: 0x9099c30
   52:   safemalloc.c:   128: _mymalloc: enter: Size: 16384
   53:   safemalloc.c:   197: _mymalloc: exit: ptr: 0x909cc80
   54:   safemalloc.c:   262: _myfree: enter: ptr: 0x9095bf8
   55:   sql_parse.cc:  1757: dispatch_command: query: SHOW AUTHORS
..
   88:    sql_show.cc:   385: mysqld_show_authors: packet_header: Memory: 0x940590  Bytes: (4)
..
  166:    protocol.cc:   385: send_eof: packet_header: Memory: 0x940550  Bytes: (4)
05 00 00 51
  167:    viosocket.c:   184: vio_is_blocking: exit: 0
  168:    viosocket.c:   105: vio_write: enter: sd: 23, buf: 0x0x9091bd0, size: 5107
  169:    viosocket.c:   117: vio_write: exit: 5107
  170:    protocol.cc:   342: send_eof: info: EOF sent, so no more error sending allowed
  171:     sql_lex.cc:   199: lex_end: enter: lex: 0x9068b58
  172:   sql_parse.cc:  1796: dispatch_command: info: query ready
  173:     my_alloc.c:   331: free_root: enter: root: 0x9068b38  flags: 1
  174:    viosocket.c:   184: vio_is_blocking: exit: 0
  175:    viosocket.c:    36: vio_read: enter: sd: 23, buf: 0x0x9091bd0, size: 4
  176:    viosocket.c:    49: vio_read: vio_error: Got error 11 during read
  177:    viosocket.c:    52: vio_read: exit: -1
  178:   sql_parse.cc:   809: do_command: info: vio_read returned -1,  errno: 11

  190:   sql_parse.cc:  1536: do_command: info: Command on socket (23) = 1 (Quit)
  191:     my_alloc.c:   331: free_root: enter: root: 0x9068b38  flags: 1
  192:      mysqld.cc:  1681: close_connection: enter: fd: socket (23)  error: ''
  193: sql_handler.cc:   628: mysql_ha_flush: enter: tables: (nil)  mode_flags: 0x02

This will make a lot more sense with my upcoming presentation where you will see clear use and interaction between source files such as sql_parse.cc & sql_show.cc.

For more information check out Making Trace Files

Compiling MySQL Tutorial 2 – Directly from the source


Should you want to be on the bleeding edge, or in my case, don’t want to download 70MB each day in a daily snapshot (especially when I’m getting build errors), you can use Bit Keeper Free Bit Keeper Client that at least lets you download the MySQL Repository. This client doesn’t allow commits, which is a good thing for those non-gurus in mysql internals (which definitely includes me).

wget http://www.bitmover.com/bk-client.shar
/bin/sh bk-client.shar
cd bk_client-1.1
make

By placing sfioball in your path you can execute.

sfioball bk://mysql.bkbits.net/mysql-5.1 mysql-5.1

This took me about 4 mins, which seemed much quicker then getting a snapshot!

You can then get cracking with my instructions at Compiling MySQL Tutorial 1 – The Baseline.

A good reference in all this compiling is to take a good look at the MySQL Internals Manual. (which I only found out about recently)

If at a later time you want to update your repository to the latest, use the following command.

update bk://mysql.bkbits.net/mysql-5.1 mysql-5.1

Documentation References

5.1 Reference Manual – 2.9. MySQL Installation Using a Source Distribution
5.1 Reference Manual – 2.9.3. Installing from the Development Source Tree
5.1 Reference Manual – 2.13.1.3. Linux Source Distribution Notes
MySQL Internals Manual

Requirements for compiling

To confirm earlier notes on minimum requirements for compiling the following details should be confirmed.

automake --version
autoconf --version
libtool --version
m4 --version
gcc --version
gmake --version
bison -version


My results running Fedora Core 5.

$ automake --version
automake (GNU automake) 1.9.6
$ autoconf --version
autoconf (GNU Autoconf) 2.59
$ libtool --version
ltmain.sh (GNU libtool) 1.5.22 (1.1220.2.365 2005/12/18 22:14:06)
$ m4 --version
GNU M4 1.4.4
$ gcc --version
gcc (GCC) 4.1.1 20060525 (Red Hat 4.1.1-1)
$ gmake --version
GNU Make 3.80
$ bison -version
bison (GNU Bison) 2.1

Compiling MySQL Tutorial 1 – The Baseline – Update

Just to confirm my earlier confusion about verified snapshots at Compiling MySQL Tutorial 1 – The Baseline.

“Daily snapshot sources are only published, if they compiled successfully (using the BUILD/compile-pentium-debug-max script) and passed the test suite (using make test). If the source tree snapshot fails to compile or does not pass the test suite, the source tarball will not be published.”

Seems the fine print at MySQL Database Server 5.1: Beta snapshots also states this. Well, need to take my RTFM pill there.

Thanks to Lenz for putting the record straight, and helping with my Forum Post. Seems I did uncover a Bug, now recorded as Bug #21463. Just need to get it fixed to continue on my tutorial.

Compiling MySQL Tutorial 1 – The Baseline

Pre-requisites

This tutorial is aimed at Linux installations that has the standard development tools already installed. The INSTALL file in the source archives provides good details of the software required (e.g. gunzip, gcc 2.95.2+,make).

Create a separate user for this development work.

su -
useradd mysqldev
su - mysqldev

Get an appropiate Snapshot

You can view recent snapshots at http://downloads.mysql.com/snapshots/mysql-5.1/.

The official statement on snapshots from MySQL AB.
Daily snapshot sources are only published, if they compiled successfully (using the BUILD/compile-pentium-debug-max script) and passed the test suite (using make test). If the source tree snapshot fails to compile or does not pass the test suite, the source tarball will not be published.

At the time of producing this the present snapshot was http://downloads.mysql.com/snapshots/mysql-5.1/mysql-5.1.12-beta-nightly-20060801.tar.gz

mkdir -p $HOME/src
cd $HOME/src
SNAPSHOT="mysql-5.1.12-beta-nightly-20060801";export SNAPSHOT
wget http://downloads.mysql.com/snapshots/mysql-5.1/${SNAPSHOT}.tar.gz
tar xfz ${SNAPSHOT}.tar.gz
cd $SNAPSHOT

Compiling

You can for example do the simple way with most C programs:

DEPLOY=${HOME}/deploy; export DEPLOY
./configure --prefix=${DEPLOY}
make
make install

However it’s recommended you use the significant number of pre-configured build scripts found in the BUILD directory. For Example:

./BUILD/compile-pentium-debug --prefix=${DEPLOY}

I always get the following warnings. Not sure what to do about it, but it doesn’t break any future work to date. Surely somebody in the development team knows about them.

../build_win32/include.tcl: no such file or directory
cp: cannot create regular file `../test/logtrack.list': No such file or directory

Testing

The easy, but time consuming part over, let’s test this new beast.

make install

This will deploy to the preconfigured area of $USER/deploy. For multiple versions within this process you should adjust this back in the compile process.

cd $DEPLOY
bin/mysql_install_db
bin/mysqld_safe &
bin/mysql -e "SELECT VERSION()"
bin/mysqladmin -uroot shutdown

Changes

So if you haven’t already sniffed around there are some reasonable changes in the structure. Here are a few points that caught me out:

  • mysql_install_db is now under /bin not /scripts, infact there is no /scripts
  • mysqld is not under /bin but infact under /libexec. A good reason to always used mysqld_safe
  • by default the data directory is /var, normal documentation etc always has this as /data

This is a quick confirm it all works, and I’ve for the purposes of this example ensured that no other MySQL installation is running, and there is no default /etc/my.cnf file. (I always place this in the MySQL installation directory anyway).

Some minonr considerations for improvements with locations, users and multiple instances.

cd $DEPLOY
rm -rf var
bin/mysql_install_db --datadir=${DEPLOY}/data --user=$USER
bin/mysqld_safe --basedir=${DEPLOY} --datadir=${DEPLOY}/data --user=$USER --port=3307 --socket=/tmp/mysqldev.sock &
bin/mysql -P3307 -S/tmp/mysqldev.sock -e "SELECT VERSION()"
bin/mysqladmin -P3307 -S/tmp/mysqldev.sock  -uroot shutdown

Troubleshooting

Now, there is no guarantee that the snapshot is a working one, in the case of mysql-5.1.12-beta-nightly-20060801 in this example, it didn’t work for me? I’ve just reinstalled my OS to Fedora Core 5, and the previous source version (a 5.1.10 snapshot worked ok)
Here are some tips to help out.

The preconfigured BUILD scripts have a nice display option with -n that allows you to see what will happen.

./BUILD/compile-pentium-debug --prefix=${DEPLOY} -n

In my case it gave me this:

testing pentium4 ... ok
gmake -k distclean || true
/bin/rm -rf */.deps/*.P config.cache storage/innobase/config.cache
   storage/bdb/build_unix/config.cache bdb/dist/autom4te.cache autom4te.cache innobase/autom4te.cache;

path=./BUILD
. "./BUILD/autorun.sh"
CC="gcc" CFLAGS="-Wimplicit -Wreturn-type -Wswitch -Wtrigraphs -Wcomment -W -Wchar-subscripts -Wformat -Wparentheses -Wsign-compare -Wwrite-strings -Wunused
-DUNIV_MUST_NOT_INLINE -DEXTRA_DEBUG -DFORCE_INIT_OF_VARS  -DSAFEMALLOC -DPEDANTIC_SAFEMALLOC -DSAFE_MUTEX"
CXX="gcc" CXXFLAGS="-Wimplicit -Wreturn-type -Wswitch -Wtrigraphs -Wcomment -W -Wchar-subscripts -Wformat -Wparentheses -Wsign-compare -Wwrite-strings -Wctor-dtor-privacy
-Wnon-virtual-dtor -felide-constructors -fno-exceptions -fno-rtti  -DUNIV_MUST_NOT_INLINE -DEXTRA_DEBUG -DFORCE_INIT_OF_VARS
-DSAFEMALLOC -DPEDANTIC_SAFEMALLOC -DSAFE_MUTEX" CXXLDFLAGS=""
./configure --prefix=/home/mysqldev/deploy --enable-assembler  --with-extra-charsets=complex  --enable-thread-safe-client
--with-readline  --with-big-tables  --with-debug=full --enable-local-infile
gmake -j 4

In my case this helped with my problem, that is the autorun.sh command is throwing an error in the Innodb configuration.
Part of the compiling is you can turn Innodb off with –without-innodb, but the BUILD scripts don’t accept parameters, so it’s a matter of ignoring the autorun.sh command above, and adding –without-innodb to the configure.

As it turned out, the autorun.sh provided the additional storage engine support.

Details of the problem in this instance.

[mysqldev@lamda mysql-5.1.12-beta-nightly-20060801]$ . "./BUILD/autorun.sh"
Updating Berkeley DB source tree permissions...
../build_win32/include.tcl: no such file or directory
Updating Berkeley DB README file...
Building ../README
autoconf: building aclocal.m4...
autoconf: running autoheader to build config.hin...
autoconf: running autoconf to build configure
Building ../test/logtrack.list
cp: cannot create regular file `../test/logtrack.list': No such file or directory
Building ../build_win32/db.h
configure.in:723: required file `zlib/Makefile.in' not found
Can't execute automake

The good news is today, the MySQL Barcamp was announced, a great place to get some better insite into the gurus of the MySQL internals. At least a place to ask these questions providing they have a beginner group.

My Next Tutorial will take a look at the example provided by Brian Aker at the recent MySQL Users Conference HackFest B: Creating New SHOW Commands.

Hacking MySQL Source improvements

Further to my earlier post Hacking MySQL Source (in a good way) in which I was having a compilation problem when re-compiling MySQL source (i.e. tt worked the first build, but failed on subsequent re-compiles, even with no changes).
I’ve been able to solve my re-compiling issue, with special thanks to Jan Kenschke of MySQL AB, who was near during the MySQL Quiz Show.

As suspected, there is no need to re-run the ./BUILD/compile-pentium-debug command as this cleans, a simple make command is sufficient (I needed to confirm no other command args where needed). By doing this, it also better highlighted the actual error, where previously is wasn’t. I didn’t keep a copy of the error, but in effect I needed to do the following in order to successfully recompile.

su - mysqldev
cd mysql-5.1.10-beta-nightly-20060413
touch mysql-test/std_data/client-key.pem
touch mysql-test/std_data/client-cert.pem
touch mysql-test/std_data/cacert.pem
touch mysql-test/std_data/server-cert.pem
touch mysql-test/std_data/server-key.pem
make
# Success

I can’t answer why I had to do this manually (create these cert permissions), and why some level of make doesn’t (or does but I don’t know the command), but it works, and right now that’s important. Hopefully a MySQL guru can enlightment me.

From here, I was able to test as I’d written previously. I just need to confirm my changes work before reporting back, as well as providing some guidelines for debugging with gdb.

Hacking MySQL Source (in a good way)

HackFest B: Creating a New SHOW Command by Brian Aker at the MySQL Users Conference

Brian stepped through the steps for those attending to modify and deploy new functionality in the mysql server. Cool. The end result I would consider for an experienced developer as relatively easy (after avoiding the pitfalls).

NOTE: I wasn’t able to complete this successfully during the session, but I’ve posted this, so hopefully the input and review of others can help in overcoming the current issues. See Outstanding Issues throughout my notes.

Our goal: Produce a new command SHOW CONFERENCE;

The following commands were performed on CentOS 4.2. There may be some differences with different Linux Distros.

Prepare a current Source Code Tree


su -
useradd mysqldev
su - mysqldev
wget http://downloads.mysql.com/snapshots/mysql-5.1/mysql-5.1.10-beta-nightly-20060426.tar.gz
# NOTE: You should check the snapshots page for latest versions http://downloads.mysql.com/snapshots.php
tar xvfz mysql-5.1.10-beta-nightly-20060413.tar.gz
cd mysql-5.1.10-beta-nightly-20060413
# compile script will depend on H/W and requirements
./BUILD/compile-pentium-debug --prefix=/home/mysqldev
make install
scripts/mysql_install_db --datadir=/home/mysqldev/data
sql/mysqld --basedir=/home/mysqldev --datadir=/home/mysqldev/data &

NOTE: This can take a while, so it’s a good opportunity to leave this and come back at a later time. The –prefix allows you to install into the new users directory ensuring that you can test without affecting any currently installed mysql installations on your machine.

The BUILD directory contains many different compilations for platforms and variants. This version contains 41 specific scripts. For the purpose of our new SHOW command, we are going to run in debug mode for any necessary debugging.

Outstanding Issues

  • I was surprised a make install didn’t configure the bin/mysqld and scripts/mysql_install_db as you would get in a binary distribution. There is probably additional commands necessary for this, but for the purpose of development it works.
  • I run multiple versions of MySQL on my machine, and I had some conflict with my /etc/my.cnf file, so I just removed it for the purpose of this test. There needs to be a cleaner way to ensure this isn’t used, and also to compartmentalise/isolate different source trees.

Confirm Installation

I ran this in a seperate terminal window.

su - mysqldev
bin/mysql -e "SELECT VERSION()"
+------------------------------------+
| VERSION()                          |
+------------------------------------+
| 5.1.10-beta-nightly-20060413-debug |
+------------------------------------+

This gives us a suitable source baseline.

Implementing a new SHOW Command

The easiest means of developing a new SHOW command is to base this on an existing command. We are going to base this new SHOW command on the SHOW AUTHORS command.

We are going to be looking at the following files.

  1. lex.h
  2. sql_yacc.yy
  3. sql_lex.h
  4. sql_parse.cc
  5. mysql_priv.h
  6. sql_show.cc

All these files are found under the sql directory in the source tree.

lex.h

131:     { "CONFERENCE",       SYM(CONFERENCE_SYM)},

sql_yacc.yy

 202:  %token  CONFERENCE_SYM
...
8267:       | CONFERENCE_SYM
8268:         {
8269:           LEX *lex=Lex;
8270:           lex->sql_command= SQLCOM_SHOW_CONFERENCE;
8271:         }
...
9366:       | CONFERENCE_SYM        {}

sql_lex.h

112:   SQLCOM_SHOW_CONFERENCE,

sql_parse.cc

3515:   case SQLCOM_SHOW_CONFERENCE:
3516:     res= mysqld_show_conference(thd);
3517:     break;

mysql_priv.h

915: bool mysqld_show_conference(THD *thd);

sql_show.cc

229: bool mysqld_show_conference(THD *thd)
230: {
231:   List<item> field_list;
232:   Protocol *protocol= thd->protocol;
233:   DBUG_ENTER("mysqld_show_conference");
234:
235:   field_list.push_back(new Item_empty_string("Name",100));
236:
237:   if (protocol->send_fields(&field_list,
238:                             Protocol::SEND_NUM_ROWS | Protocol::SEND_EOF))
239:     DBUG_RETURN(TRUE);
240:
241:   protocol->prepare_for_resend();
242:   protocol->store("Welcome to the MySQL User Conference 2006", system_charset_info);
243:   if (protocol->write())
244:     DBUG_RETURN(TRUE);
245:   send_eof(thd);
246:   DBUG_RETURN(FALSE);
247: }


cd ..
./BUILD/compile-pentium-debug --prefix=/home/mysqldev
...
gmake[3]: *** Waiting for unfinished jobs....
/bin/chmod +x mysql-test-run-t
/bin/mv mysql-test-run-t mysql-test-run
gmake[3]: Leaving directory `/home/mysqldev/mysql-5.1.10-beta-nightly-20060413/mysql-test'
gmake[2]: *** [all-recursive] Error 1
gmake[2]: Leaving directory `/home/mysqldev/mysql-5.1.10-beta-nightly-20060413/mysql-test'
gmake[1]: *** [all-recursive] Error 1
gmake[1]: Leaving directory `/home/mysqldev/mysql-5.1.10-beta-nightly-20060413'
gmake: *** [all] Error 2

Outstanding Issues

  • I’m sure there must be a more effecient means of compiling changes, this process effectively cleaned and started again.
  • I had a build failure which didn’t seem to make any sense.

At this point I wasn’t able to continue, but here are notes I took of next steps.

make install
gdb mysqld
run --gdb --debug

In another terminal session.

su - mysqldev
bin/mysql
mysql> SHOW CONFERENCE;

I’m keen to see it work in my own environment and documented for others to try.

Brian moved on to creating a INFORMATION_SCHEMA query but we ran out of time to complete this. I’ve got some notes to document as a later date.