Oracle Linux/MySQL - Revision history

R00t: /* Script */

2020-03-18T23:04:24Z

‎Script

R00t: /* haproxy */

2020-03-18T22:35:21Z

‎haproxy

R00t: /* NDB cluster status script */

2020-03-18T22:33:53Z

‎NDB cluster status script

R00t at 18:10, 18 March 2020

2020-03-18T18:10:35Z

R00t: /* NDB cluster status script */

2020-03-18T14:58:45Z

‎NDB cluster status script

R00t: /* NDB cluster status script */

2020-03-17T18:10:26Z

‎NDB cluster status script

R00t: /* NDB cluster status script */

2020-03-17T18:07:39Z

‎NDB cluster status script

R00t at 18:05, 17 March 2020

2020-03-17T18:05:49Z

R00t: /* configure data node */

2020-03-12T17:11:36Z

‎configure data node

R00t: /* configure management node cluster */

2020-03-10T16:34:14Z

‎configure management node cluster

@@ Line 449: / Line 449: @@
              rm $logdir/test
          else
-             echo "$dateprefix unable to write to log directory: $logdir" &> /tmp/xinetd.log
+             echo "$dateprefix unable to write to log directory: $logdir" &>> /tmp/xinetd.log
              #exit 1
          fi
      else
-         echo "$dateprefix log directory does not exist: $logdir" &> /tmp/xinetd.log
+         echo "$dateprefix log directory does not exist: $logdir" &>> /tmp/xinetd.log
          #exit 1
      fi
@@ Line 460: / Line 460: @@
 # function to check if the specified package exists
 function check_packages {
-     echo "$dateprefix Checking Packages" &> $logfile
+     echo "$dateprefix Checking Packages" &>> $logfile
-     packages=("xinetd-2.3.15-13.el7.x86_64" "nmap.x86_64 2:6.40-19.el7")
+     packages=("xinetd" "nmap")
      for package in ${packages[@]}
      do
          rpm -qa | grep $package &>/dev/null
          if [ $? -ne 0 ]; then
-             echo "$dateprefix package: $package is not installed.  EOL" &> $logfile
+             echo "$dateprefix package: $package is not installed.  EOL" &>> $logfile
              #exit 1
          fi
@@ Line 475: / Line 475: @@
 # return a 503 HTTP response if either condition is false, otherwise return a 200 HTTP response.
 function check_sql {
-     echo "$dateprefix Checking SQL" &> $logfile
+     echo "$dateprefix Checking SQL" &>> $logfile
      #change default field separator, white space, to newline character so arrays are parsed properly
      saveifs=$IFS
@@ Line 484: / Line 484: @@
          "Content-Length: 26\r\n"
          "\r\n"
-         "SQL server is available.\r\n")
+         "SQL server is available.\r\n"
      declare -a no_datanodes=("HTTP/1.1 503 Service Unavailable\r\n"
          "Content-Type: text/plain\r\n"
@@ Line 490: / Line 491: @@
          "Content-Length: 40\r\n"
          "\r\n"
-         "No available data nodes from SQL node.\r\n")
+         "No available data nodes from SQL node.\r\n"
      declare -a sql_unavailable=("HTTP/1.1 503 Service Unavailable\r\n"
          "Content-Type: text/plain\r\n"
@@ Line 496: / Line 498: @@
          "Content-Length: 27\r\n"
          "\r\n"
-         "SQL Node is unresponsive.\r\n")
+         "SQL Node is unresponsive.\r\n"
      nc -z $sqlhost 3306
      if [ $? -eq 0 ]; then

@@ Line 554: / Line 554: @@
 example config
   <nowiki>listen mysql-cluster 0.0.0.0:3306
-	mode http
+	mode tcp
 	balance roundrobin
 	option httpchk GET /
 	server sql01 rncv-sql01.virtual.local:3306 check port 5900

← Older revision		Revision as of 22:33, 18 March 2020
Line 559:		Line 559:
	server sql01 rncv-sql01.virtual.local:3306 check port 5900		server sql01 rncv-sql01.virtual.local:3306 check port 5900
	server sql02 rncv-sql02.virtual.local:3306 check port 5900</nowiki>		server sql02 rncv-sql02.virtual.local:3306 check port 5900</nowiki>
		+
		+	==mysql access==
		+	Keep in mind that you'll need to either create a user specifically for haproxy with the host defined as the name/ip, or specify the host as <code>%</code> so that the user can log in from anywhere.
		+	<nowiki> mysql> create user `admin`@`%` identified by '^R5^2Jzz@$La-]x[H;3}';
		+	Query OK, 0 rows affected (0.02 sec)
		+
		+	mysql> grant all privileges on . to `admin`@`%` with grant option;
		+	Query OK, 0 rows affected (0.33 sec)</nowiki>
		+	In the case that you've tried to log into MySQL too many time through the haproxy, it could cause haproxy to get locked out. flush the hosts from mysql to reset.
		+	<nowiki>mysql --user='root' --password='7,q(hf`U+3=?STg4]5a4xG!7DQL%Le' -sN -e "flush hosts;"</nowiki>

@@ Line 482: / Line 482: @@
          "Content-Type: text/plain\r\n"
          "Connection: close\r\n"
-         "Content-Length: 40\r\n"
+         "Content-Length: 26\r\n"
          "\r\n"
          "SQL server is available.\r\n")
@@ Line 488: / Line 488: @@
          "Content-Type: text/plain\r\n"
          "Connection: close\r\n"
-         "Content-Length: 44\r\n"
+         "Content-Length: 40\r\n"
          "\r\n"
          "No available data nodes from SQL node.\r\n")
@@ Line 494: / Line 494: @@
          "Content-Type: text/plain\r\n"
          "Connection: close\r\n"
-         "Content-Length: 44\r\n"
+         "Content-Length: 27\r\n"
          "\r\n"
          "SQL Node is unresponsive.\r\n")
@@ Line 550: / Line 550: @@
 ~$ sudo firewall-cmd --reload
 success</nowiki>

← Older revision		Revision as of 14:58, 18 March 2020
Line 544:		Line 544:
	type = UNLISTED		type = UNLISTED
	}</nowiki>		}</nowiki>
		+
		+	==firewalld rules==
		+	<nowiki>~$ sudo firewall-cmd --zone=internal --add-port=5900/tcp --permanent
		+	success
		+	~$ sudo firewall-cmd --reload
		+	success</nowiki>

@@ Line 404: / Line 404: @@
 #This in turn made it impossible for HAProxy's built in mysql status checking plugin to work.
 #Something else i'm forgetting as well.
-So, in lieu of these issues, I opted to write my own NDB cluster status service that the HAProxy httpd plugin can query to determine availability.  [https://raw.githubusercontent.com/olafz/percona-clustercheck/master/clustercheck This is based off of this script].
+So, in lieu of these issues, I opted to write my own NDB cluster status service that the HAProxy httpd plugin can query to determine availability.  [https://raw.githubusercontent.com/olafz/percona-clustercheck/master/clustercheck This is based off of this script].<br>
 <div class="toccolours mw-collapsible mw-collapsed">
 NDB cluster status script:
@@ Line 529: / Line 531: @@
 check_sql
 </nowiki></pre></div></div>

@@ Line 409: / Line 409: @@
 NDB cluster status script:
 <div class="mw-collapsible-content"><pre style="white-space: pre-wrap;"><nowiki>
-Test
+#!/bin/bash
 </nowiki></pre></div></div>

← Older revision		Revision as of 18:05, 17 March 2020
Line 398:		Line 398:
	You should only need to restart the mysqld.service on the rest of the SQL nodes for them to pull in the changes.		You should only need to restart the mysqld.service on the rest of the SQL nodes for them to pull in the changes.
	<nowiki>Whenever a MySQL server starts up and joins the cluster as an SQL node it executes these stored CREATE USER and GRANT statements as part of the cluster schema synchronization process.</nowiki>		<nowiki>Whenever a MySQL server starts up and joins the cluster as an SQL node it executes these stored CREATE USER and GRANT statements as part of the cluster schema synchronization process.</nowiki>
		+
		+	=NDB cluster status script=
		+	Using HAProxy in front of a MySQL 8 NDB cluster, i ran into a couple of challenges.
		+	#I could not get user/pass data to sync across the mysql nodes
		+	#This in turn made it impossible for HAProxy's built in mysql status checking plugin to work.
		+	#Something else i'm forgetting as well.
		+	So, in lieu of these issues, I opted to write my own NDB cluster status service that the HAProxy httpd plugin can query to determine availability. [https://raw.githubusercontent.com/olafz/percona-clustercheck/master/clustercheck This is based off of this script].
		+
		+	<div class="toccolours mw-collapsible mw-collapsed">
		+	NDB cluster status script:
		+	<div class="mw-collapsible-content"><pre style="white-space: pre-wrap;"><nowiki>
		+	Test
		+	</nowiki></pre></div></div>

@@ Line 243: / Line 243: @@
 	class file lock;
 	class file getattr;
+}
@@ Line 252: / Line 254: @@
 allow mysqld_t var_lib_t:file { read write };
 allow mysqld_t var_lib_t:file lock;
-allow mysqld_t var_lib_t:file getattr;</nowiki>
+allow mysqld_t var_lib_t:file getattr;
 ==start data node==

@@ Line 41: / Line 41: @@
 ==configure management node cluster==
-Now we can setup a basic configuration file for the cluster.
+Now we can setup a basic configuration file for the cluster. [https://dev.mysql.com/doc/mysql-cluster-excerpt/8.0/en/mysql-cluster-config-starting.html reccommended]
   <nowiki>~$ sudo mkdir -p /var/lib/mysql-cluster
 ~$ sudo vim /var/lib/mysql-cluster/config.ini