Comments on: A quest for the full InnoDB status http://www.mysqlperformanceblog.com/2008/10/31/full-innodb-status/ Everything about MySQL Performance Sat, 21 Nov 2009 14:06:19 -0800 http://wordpress.org/?v=2.8.4 hourly 1 By: Arjen Lentz http://www.mysqlperformanceblog.com/2008/10/31/full-innodb-status/comment-page-1/#comment-370801 Arjen Lentz Sat, 01 Nov 2008 23:43:01 +0000 http://www.mysqlperformanceblog.com/?p=519#comment-370801 Ok, so why don't we just take out the limiter. If people ask for it, they want it. Indeed we can do this inside Baron's patch so people can still control what they want to see. Ok, so why don’t we just take out the limiter. If people ask for it, they want it. Indeed we can do this inside Baron’s patch so people can still control what they want to see.

]]> By: Maciej Dobrzanski http://www.mysqlperformanceblog.com/2008/10/31/full-innodb-status/comment-page-1/#comment-370657 Maciej Dobrzanski Sat, 01 Nov 2008 16:48:28 +0000 http://www.mysqlperformanceblog.com/?p=519#comment-370657 Arjen, It's definietly a server-side thing. All is done in ha_innodb.cc:innodb_show_status(). Maciek Arjen,

It’s definietly a server-side thing. All is done in ha_innodb.cc:innodb_show_status().

Maciek

]]> By: Baron Schwartz http://www.mysqlperformanceblog.com/2008/10/31/full-innodb-status/comment-page-1/#comment-370459 Baron Schwartz Sat, 01 Nov 2008 13:12:18 +0000 http://www.mysqlperformanceblog.com/?p=519#comment-370459 I think it's the server. The code is in InnoDB. There are a couple places it happens; in one place it actually checks how big the output is and stops outputting any more txns. The best tweak is a) my patches, which reduce the verbosity, b) Google's patches, which increase the allowed size and move the deadlock and other less-useful stuff to the end of the output, so if it gets truncated at least you get the important stuff. I think it’s the server. The code is in InnoDB. There are a couple places it happens; in one place it actually checks how big the output is and stops outputting any more txns.

The best tweak is a) my patches, which reduce the verbosity, b) Google’s patches, which increase the allowed size and move the deadlock and other less-useful stuff to the end of the output, so if it gets truncated at least you get the important stuff.

]]> By: Arjen Lentz http://www.mysqlperformanceblog.com/2008/10/31/full-innodb-status/comment-page-1/#comment-369911 Arjen Lentz Fri, 31 Oct 2008 23:28:46 +0000 http://www.mysqlperformanceblog.com/?p=519#comment-369911 I should check again, but I believe that it's the mysql cmdline client that truncates the output, not the server. So it is possible to get it. We could tweak so that the innodb-status-file option is enabled by default... doesn't make much sense not to, if it writes to a file first anyway.... but I prefer a solution that can be accessed from the client side (although slow query log can't be either in 5.0). I should check again, but I believe that it’s the mysql cmdline client that truncates the output, not the server. So it is possible to get it.

We could tweak so that the innodb-status-file option is enabled by default… doesn’t make much sense not to, if it writes to a file first anyway…. but I prefer a solution that can be accessed from the client side (although slow query log can’t be either in 5.0).

]]>