filter: select, mysql Solaris 3.23.40 connections using perl, jdbc, odbc.
Yesterday, we experienced a rare mysqld failure where all connections where in the state="Opening table". Normally our 200+ connections are in state="sleep". The Time field from command "show processlist" showed each connection accumulating time. This is not the case when a connection is in the sleep mode. The number of connections was nearly 400 when we normally have over 200. The threads running about equal to number of connections. The threads created was at 1240 - normally is ~ number of connections. When I attempted a "mysql.server stop", it timed out. I had to use a "kill -9". I used a combination of myismcheck (db off) and check table (db on) to verify that all tables were ok. I did repair tables that were OK but terminated improperly probably due to the "kill -9" command. The info column from the "show processlist" showed the command the user or program was executing. All of the connections were accounted for including the extra 170 connections. These extra connections accumulated from programs between the time server failed and time of restart. To me, it appears that the mysqld allowed connections, but then froze with all connections executing at the state "Opening table". This is verified from programs making connections without failing. It appears all were waiting for mysqld to process the query. Just in case this matters, there are ~75 tables in this installation. On a normal day, open_tables=149 and table cache is set to 256. Prior to restarting mysqld, opened_tables was at 831. Any suggestions welcome. David -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
