Annotation of doc/build/loncapasqldatabase.html, revision 1.6

1.1       harris41    1: <HTML>
                      2: <HEAD>
                      3: <TITLE>LON-CAPA SQL Database Documentation</TITLE>
                      4: </HEAD>
                      5: <BODY>
                      6: <H1>LON-CAPA SQL Database Documentation</H1>
                      7: <P>
                      8: Scott Harrison
                      9: </P>
                     10: <P>
1.5       harris41   11: Last updated: 02/12/2001
1.1       harris41   12: </P>
1.2       harris41   13: <P>
1.1       harris41   14: This file describes issues associated with LON-CAPA
                     15: and a SQL database.
                     16: </P>
1.5       harris41   17: <H2>Latest HOWTO</H2>
1.2       harris41   18: <P>
1.4       harris41   19: <UL>
                     20: <LI>Current status of documentation</LI>
                     21: <LI>Current status of implementation</LI>
                     22: <LI>Purpose within LON-CAPA</LI>
                     23: <LI>Installation</LI>
                     24: <LI>Installation from source</LI>
                     25: <LI>Configuration (automated)</LI>
                     26: <LI>Manual configuration</LI>
                     27: <LI>Testing</LI>
                     28: <LI>Example sections of code relevant to LON-CAPA</LI>
                     29: </UL>
                     30: </P>
1.5       harris41   31: <H2>Current status of documentation</H2>
1.4       harris41   32: <P>
1.2       harris41   33: I am going to begin documentation by inserting what notes
                     34: I have into this file.  I will be subsequently rearranging
1.4       harris41   35: them and editing them based on the tests that I conduct.
1.2       harris41   36: I am trying to make sure that documentation, installation,
                     37: and run-time issues are all consistent and correct.  The
                     38: current status of everything is that it works and has
                     39: been minimally tested, but things need to be cleaned up
                     40: and checked again!
                     41: </P>
1.5       harris41   42: <H2>Current status of implementation</H2>
1.4       harris41   43: <P>
1.6     ! harris41   44: Need to
        !            45: <UL>
        !            46: <LI>Installation: Fix binary file listings for user permissions and ownership.
        !            47: <LI>Installation: Make sure sql server starts, and if database does not
        !            48: exist, then create. (/etc/rc.d).
        !            49: <LI>Processes: Make sure loncron initiates lonsql on library machines.
        !            50: <LI>Read in metadata from right place periodically.
        !            51: <LI>Implement tested perl module handler.
        !            52: </UL>
        !            53: <P>
1.4       harris41   54: Right now, a lot of "feasibility" work has been done.
                     55: Recipes for manual installation and configuration have
                     56: been gathered.  Network connectivity of lond->lonsql->lond->lonc
                     57: type tests have been performed.  A binary installation
                     58: has been compiled in an RPM (LON-CAPA-mysql).
                     59: The most lacking test in terms of feasibility has
                     60: been looking at benchmarks to analyze the load at which
                     61: the SQL database can efficiently allow many users to
                     62: make simultaneous requests of the metadata database.
                     63: </P>
                     64: <P>
                     65: Documentation has been pieced together over time.  But,
                     66: as mentioned in the previous section, it needs an
                     67: overhaul.
                     68: </P>
                     69: <P>
                     70: The binary installation has some quirks associated with it.
                     71: Some of the user permissions are wrong, although this is
                     72: benign.  Also, other options of binary installation (such
                     73: as using binary RPMs put together by others) were dismissed
                     74: given the difficulty of getting differing combinations of
                     75: these external RPMs to work together.
                     76: </P>
                     77: <P>
                     78: Most configuration questions have been initially worked out
                     79: to the point of getting this SQL software component working,
                     80: however there may be more optimal approaches than currently
                     81: exist.
                     82: </P>
1.5       harris41   83: <H2>Purpose within LON-CAPA</H2>
1.4       harris41   84: <P>
                     85: LON-CAPA is meant to distribute A LOT of educational content
                     86: to A LOT of people.  It is ineffective to directly rely on contents
                     87: within the ext2 filesystem to be speedily scanned for 
                     88: on-the-fly searches of content descriptions.  (Simply put,
                     89: it takes a cumbersome amount of time to open, read, analyze, and
                     90: close thousands of files.)
                     91: </P>
                     92: <P>
                     93: The solution is to hash-index various data fields that are
                     94: descriptive of the educational resources on a LON-CAPA server
                     95: machine.  Descriptive data fields are referred to as
                     96: "metadata".  The question then arises as to how this metadata
                     97: is handled in terms of the rest of the LON-CAPA network
                     98: without burdening client and daemon processes.  I now
                     99: answer this question in the format of Problem and Solution
                    100: below.
                    101: </P>
                    102: <P>
                    103: <PRE>
                    104: PROBLEM SITUATION:
                    105: 
                    106:   If Server A wants data from Server B, Server A uses a lonc process to
                    107:   send a database command to a Server B lond process.
                    108:     lonc= loncapa client process    A-lonc= a lonc process on Server A
                    109:     lond= loncapa daemon process
                    110: 
                    111:                  database command
                    112:     A-lonc  --------TCP/IP----------------> B-lond
                    113: 
                    114:   The problem emerges that A-lonc and B-lond are kept waiting for the
                    115:   MySQL server to "do its stuff", or in other words, perform the conceivably
                    116:   sophisticated, data-intensive, time-sucking database transaction.  By tying
                    117:   up a lonc and lond process, this significantly cripples the capabilities
                    118:   of LON-CAPA servers. 
                    119: 
                    120:   While commercial databases have a variety of features that ATTEMPT to
                    121:   deal with this, freeware databases are still experimenting and exploring
                    122:   with different schemes with varying degrees of performance stability.
                    123: 
                    124: THE SOLUTION:
                    125: 
                    126:   A separate daemon process was created that B-lond works with to
                    127:   handle database requests.  This daemon process is called "lonsql".
                    128: 
                    129:   So,
                    130:                 database command
                    131:   A-lonc  ---------TCP/IP-----------------> B-lond =====> B-lonsql
                    132:          <---------------------------------/                |
                    133:            "ok, I'll get back to you..."                    |
                    134:                                                             |
                    135:                                                             /
                    136:   A-lond  <-------------------------------  B-lonc   <======
                    137:            "Guess what? I have the result!"
                    138: 
                    139:   Of course, depending on success or failure, the messages may vary,
                    140:   but the principle remains the same where a separate pool of children
                    141:   processes (lonsql's) handle the MySQL database manipulations.
                    142: </PRE>
                    143: </P>
1.5       harris41  144: <H2>Installation</H2>
1.4       harris41  145: <P>
                    146: Installation of the LON-CAPA SQL database normally occurs
                    147: by default when using the LON-CAPA installation CD
                    148: (see http://install.lon-capa.org).  It is installed
                    149: as the LON-CAPA-mysql RPM.  This RPM encodes for the MySQL
                    150: engine and related perl interfaces (Perl::DBI, Perl::Msql-Mysql).
                    151: </P>
                    152: <P>
                    153: The three components of a MySQL installation for the
                    154: LON-CAPA system are further described immediately below.
                    155: <TABLE BORDER="0">
                    156: <TR><TD COLSPAN="2"><STRONG>Perl::DBI module</STRONG>-
                    157: the API "front-end"...</TD></TR>
                    158: <TR><TD WIDTH="10%"></TD><TD>database interface module for organizing generic
                    159: database commands which are independent of specific
                    160: database implementation (such as MySQL, mSQL, Postgres, etc).
                    161: </TD></TR>
                    162: <TR><TD COLSPAN="2"><STRONG>Perl::MySQL module</STRONG>-
                    163: the API "mid-section"...</TD></TR>
                    164: <TR><TD WIDTH="10%"></TD><TD>the module to directly interface with the actual
                    165: MySQL database engine</TD></TR>
                    166: <TR><TD COLSPAN="2"><STRONG>MySQL database engine</STRONG>-
                    167: the "back-end"...</TD></TR>
                    168: <TR><TD WIDTH="10%"></TD><TD>the binary installation (compiled either
                    169: from source or pre-compiled file listings) which provides the
                    170: actual MySQL functionality on the system</TD></TR>
                    171: </TABLE>
                    172: </P>
1.5       harris41  173: <H2>Installation from source</H2>
1.4       harris41  174: <P>
                    175: The following set of tarballs was found to work together
                    176: properly on a LON-CAPA RedHat 6.2 system:
                    177: <UL>
                    178: <LI>DBI-1.13.tar.gz
                    179: <LI>Msql-Mysql-modules-1.2209.tar.gz
                    180: <LI>mysql-3.22.32.tar.gz
                    181: </UL>
                    182: </P>
                    183: <P>
                    184: Installation was simply a matter of following the instructions
                    185: and typing the several "make" commands for each 
                    186: </P>
1.5       harris41  187: <H2>Configuration (automated)</H2>
1.4       harris41  188: <P>
                    189: Not yet developed.  This will be part of an interface
                    190: present on LON-CAPA systems that can be launched by
                    191: entering the command <TT>/usr/sbin/loncapaconfig</TT>.
                    192: </P>
1.5       harris41  193: <H2>Manual configuration</H2>
1.4       harris41  194: <P>
                    195: This is not complete.
                    196: </P>
                    197: <P>
                    198: <STRONG>Starting the mysql daemon</STRONG>: Login on the Linux
                    199: system as user 'www'.  Enter the command
                    200: <TT>/usr/local/bin/safe_mysqld &</TT>
                    201: </P>
                    202: <P>
                    203: <STRONG>Set a password for 'root'</STRONG>:
                    204: <TT>/usr/local/bin/mysqladmin -u root password 'new-password'</TT>
                    205: </P>
                    206: <P>
                    207: <STRONG>Adding a user</STRONG>:  Start the mysql daemon.  Login to the
                    208: mysql system as root (<TT>mysql -u root -p mysql</TT>)
                    209: and enter the right password (for instance 'newmysql').  Add the user
                    210: www
                    211: <PRE>
                    212: INSERT INTO user (Host, User, Password)
                    213: VALUES ('localhost','www',password('newmysql'));
                    214: </PRE>
                    215: </P>
                    216: <P>
                    217: <STRONG>Granting privileges to user 'www'</STRONG>:
                    218: <PRE>
                    219: GRANT ALL PRIVILEGES ON *.* TO www@localhost;
                    220: FLUSH PRIVILEGES;
                    221: </PRE>
                    222: </P>
                    223: <P>
                    224: <STRONG>Set the SQL server to start upon system startup</STRONG>:
                    225: Copy support-files/mysql.server to the right place on the system
                    226: (/etc/rc.d/...).
                    227: </P>
1.1       harris41  228: <P>
1.5       harris41  229: <STRONG>The Perl API</STRONG>
1.2       harris41  230: <PRE>
                    231:    $dbh = DBI->connect(	"DBI:mysql:loncapa",
                    232: 			"www",
                    233: 			"SOMEPASSWORD",
                    234: 			{ RaiseError =>0,PrintError=>0});
                    235: 
                    236: There is an obvious need to CONNECT to the database, and in order to do
                    237: this, there must be:
                    238:   a RUNNING mysql daemon;
                    239:   a DATABASE named "loncapa";
                    240:   a USER named "www";
                    241:   and an ABILITY for LON-CAPA on one machine to access
                    242:        SQL database on another machine;
                    243:   
                    244: So, here are some notes on implementing these configurations.
                    245: 
                    246: ** RUNNING mysql daemon (safe_mysqld method)
                    247: 
                    248: The recommended way to run the MySQL daemon is as a non-root user
                    249: (probably www)...
                    250: 
                    251: so, 1) login as user www on the linux machine
                    252:     2) start the mysql daemon as /usr/local/bin/safe_mysqld &
                    253: 
                    254: safe_mysqld only works if the local installation of MySQL is set to the
                    255: right directory permissions which I found to be:
                    256: chown www:users /usr/local/var/mysql
                    257: chown www:users /usr/local/lib/mysql
                    258: chown -R www:users /usr/local/mysql
                    259: chown www:users /usr/local/include/mysql
                    260: chown www:users /usr/local/var
                    261: 
                    262: ** DATABASE named "loncapa"
                    263: 
                    264: As user www, run this command
                    265:     mysql -u root -p mysql
                    266: enter the password as SOMEPASSWORD
                    267: 
                    268: This allows you to manually enter MySQL commands.
                    269: The MySQL command to generate the loncapa DATABASE is:
                    270: 
                    271: CREATE DATABASE 'loncapa';
                    272: 
                    273: ** USER named "www"
                    274: 
                    275: As user www, run this command
                    276:     mysql -u root -p mysql
                    277: enter the password as SOMEPASSWORD
                    278: 
                    279: To add the user www to the MySQL server, and grant all
                    280: privileges on *.* to www@localhost identified by 'SOMEPASSWORD'
                    281: with grant option;
                    282: 
                    283: INSERT INTO user (Host, User, Password)
                    284: VALUES ('localhost','www',password('SOMEPASSWORD'));
                    285: 
                    286: GRANT ALL PRIVILEGES ON *.* TO www@localhost;
                    287: 
                    288: FLUSH PRIVILEGES;
                    289: 
                    290: ** ABILITY for LON-CAPA machines to communicate with SQL databases on
                    291:    other LON-CAPA machines
                    292: 
1.5       harris41  293: An up-to-date lond and lonsql.
                    294: </PRE>
                    295: </P>
                    296: <H2>Testing</H2>
                    297: <P>
                    298: <PRE>
                    299: <STRONG>** TEST the database connection with my current tester.pl code
                    300: which mimics what command will eventually be sent through lonc.</STRONG>
1.2       harris41  301: 
1.5       harris41  302: $reply=reply(
                    303:     "querysend:SELECT * FROM general_information WHERE Id='AAAAA'",$lonID);
                    304: </PRE>
                    305: </P>
                    306: <H2>Example sections of code relevant to LON-CAPA</H2>
                    307: <P>
1.2       harris41  308: Here are excerpts of code which implement the above handling:
1.5       harris41  309: </P>
                    310: <P>
                    311: <PRE>
                    312: <STRONG>**LONSQL
1.2       harris41  313: A subroutine from "lonsql" which establishes a child process for handling
1.5       harris41  314: database interactions.</STRONG>
1.2       harris41  315: 
                    316: sub make_new_child {
                    317:     my $pid;
                    318:     my $sigset;
                    319:     
                    320:     # block signal for fork
                    321:     $sigset = POSIX::SigSet->new(SIGINT);
                    322:     sigprocmask(SIG_BLOCK, $sigset)
                    323:         or die "Can't block SIGINT for fork: $!\n";
                    324:     
                    325:     die "fork: $!" unless defined ($pid = fork);
                    326:     
                    327:     if ($pid) {
                    328:         # Parent records the child's birth and returns.
                    329:         sigprocmask(SIG_UNBLOCK, $sigset)
                    330:             or die "Can't unblock SIGINT for fork: $!\n";
                    331:         $children{$pid} = 1;
                    332:         $children++;
                    333:         return;
                    334:     } else {
                    335:         # Child can *not* return from this subroutine.
                    336:         $SIG{INT} = 'DEFAULT';      # make SIGINT kill us as it did before
                    337:     
                    338:         # unblock signals
                    339:         sigprocmask(SIG_UNBLOCK, $sigset)
                    340:             or die "Can't unblock SIGINT for fork: $!\n";
                    341: 	
                    342: 	
                    343:         #open database handle
                    344: 	# making dbh global to avoid garbage collector
                    345: 	unless (
                    346: 		$dbh = DBI->connect("DBI:mysql:loncapa","www","SOMEPASSWORD",{ RaiseError =>0,PrintError=>0})
                    347: 		) { 
                    348: 	            my $st=120+int(rand(240));
                    349: 		    &logthis("<font color=blue>WARNING: Couldn't connect to database  ($st secs): $@</font>");
                    350: 		    print "database handle error\n";
                    351: 		    sleep($st);
                    352: 		    exit;
                    353: 
                    354: 	  };
                    355: 	# make sure that a database disconnection occurs with ending kill signals
                    356: 	$SIG{TERM}=$SIG{INT}=$SIG{QUIT}=$SIG{__DIE__}=\&DISCONNECT;
                    357: 
                    358:         # handle connections until we've reached $MAX_CLIENTS_PER_CHILD
                    359:         for ($i=0; $i < $MAX_CLIENTS_PER_CHILD; $i++) {
                    360:             $client = $server->accept()     or last;
                    361:             
                    362:             # do something with the connection
                    363: 	    $run = $run+1;
                    364: 	    my $userinput = <$client>;
                    365: 	    chomp($userinput);
                    366: 	    	    
                    367: 	    my ($conserver,$querytmp)=split(/&/,$userinput);
                    368: 	    my $query=unescape($querytmp);
                    369: 
                    370:             #send query id which is pid_unixdatetime_runningcounter
                    371: 	    $queryid = $thisserver;
                    372: 	    $queryid .="_".($$)."_";
                    373: 	    $queryid .= time."_";
                    374: 	    $queryid .= $run;
                    375: 	    print $client "$queryid\n";
                    376: 	    
                    377:             #prepare and execute the query
                    378: 	    my $sth = $dbh->prepare($query);
                    379: 	    my $result;
                    380: 	    unless ($sth->execute())
                    381: 	    {
                    382: 		&logthis("<font color=blue>WARNING: Could not retrieve from database: $@</font>");
                    383: 		$result="";
                    384: 	    }
                    385: 	    else {
                    386: 		my $r1=$sth->fetchall_arrayref;
                    387: 		my @r2; map {my $a=$_; my @b=map {escape($_)} @$a; push @r2,join(",", @b)} (@$r1);
                    388: 		$result=join("&",@r2) . "\n";
                    389: 	    }
                    390:             &reply("queryreply:$queryid:$result",$conserver);
                    391: 
                    392:         }
                    393:     
                    394:         # tidy up gracefully and finish
                    395: 	
                    396:         #close the database handle
                    397: 	$dbh->disconnect
                    398: 	   or &logthis("<font color=blue>WARNING: Couldn't disconnect from database  $DBI::errstr ($st secs): $@</font>");
                    399:     
                    400:         # this exit is VERY important, otherwise the child will become
                    401:         # a producer of more and more children, forking yourself into
                    402:         # process death.
                    403:         exit;
                    404:     }
                    405: }
1.5       harris41  406: </P>
                    407: <P>
                    408: <STRONG>** LOND enabling of MySQL requests</STRONG>
                    409: <BR />
                    410: This code is part of every lond child process in the
                    411: way that it parses command request syntax sent to it
                    412: from lonc processes.  Based on the diagram above, querysend
                    413: corresponds to B-lonc sending the result of the query.
                    414: queryreply corresponds to B-lond indicating that it has
                    415: received the request and will start the database transaction
                    416: (it returns "ok" to
                    417: A-lonc ($client)).
                    418: <PRE>
1.2       harris41  419: # ------------------------------------------------------------------- querysend
                    420:                    } elsif ($userinput =~ /^querysend/) {
                    421:                        my ($cmd,$query)=split(/:/,$userinput);
                    422: 		       $query=~s/\n*$//g;
                    423:                      print $client sqlreply("$hostid{$clientip}\&$query")."\n";
                    424: # ------------------------------------------------------------------ queryreply
                    425:                    } elsif ($userinput =~ /^queryreply/) {
                    426:                        my ($cmd,$id,$reply)=split(/:/,$userinput); 
                    427: 		       my $store;
                    428:                        my $execdir=$perlvar{'lonDaemons'};
                    429:                        if ($store=IO::File->new(">$execdir/tmp/$id")) {
                    430: 			   print $store $reply;
                    431: 			   close $store;
                    432: 			   print $client "ok\n";
                    433: 		       }
                    434: 		       else {
                    435: 			   print $client "error:$!\n";
                    436: 		       }
                    437: 
1.5       harris41  438: </PRE>
1.2       harris41  439: 
1.1       harris41  440: </P>
                    441: </BODY>
1.3       harris41  442: </HTML>

FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>