<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/">
    <channel>
        <title>The InnoDB Forums - InnoDB Storage Engine</title>
        <description>Get support for using the standard built-in InnoDB storage engine distributed by MySQL from other users and from InnoDB developers themselves.</description>
        <link>http://forums.innodb.com/list.php?4</link>
        <lastBuildDate>Tue, 07 Sep 2010 14:30:42 +0300</lastBuildDate>
        <generator>Phorum 5.2.13</generator>
        <item>
            <guid>http://forums.innodb.com/read.php?4,1293,1293#msg-1293</guid>
            <title>fast add columns in mysql innodb (3 replies)</title>
            <link>http://forums.innodb.com/read.php?4,1293,1293#msg-1293</link>
            <description><![CDATA[ Do you know if is there any planning about add mysql the option to add new columns in a faster way as ORACLE or sql server does ?, that would be an option to make a better engine in mysql ,  I like the options it have today but adding columns in big tables is terrible and make changes in production is painfull each time you need to do, adding new columns over big tables is something that make you cry during hours or days....]]></description>
            <dc:creator>CesarAugusto</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Tue, 24 Aug 2010 05:56:49 +0300</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,1273,1273#msg-1273</guid>
            <title>Write a C/C++ console application that connects to a MySQL server (with InnoDB plugin enabled) and prints information about the last detected deadlock (if any) by the InnoDB Storage Engine. (no replies)</title>
            <link>http://forums.innodb.com/read.php?4,1273,1273#msg-1273</link>
            <description><![CDATA[ pls i need the code to this program as soon as possible]]></description>
            <dc:creator>Ashwini</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Wed, 04 Aug 2010 10:21:27 +0300</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,1259,1259#msg-1259</guid>
            <title>optimize locks the whole table (4 replies)</title>
            <link>http://forums.innodb.com/read.php?4,1259,1259#msg-1259</link>
            <description><![CDATA[ Background: <br />
We have an app. that is running 24 X 7 and processing about 3 million incoming data per day. We do delete data on a routine basics. So, we run optimize on the major tables once a day. <br />
<br />
Problem:<br />
When ever optimize is running it is locking the whole table and processing stops until optimize completes its job. <br />
<br />
However, Select on those seem to work okay. It is the Insert and Update is being locked. <br />
<br />
Thanks for any input you could provide. We are using MySql 5.1.42 with InnoDB as the storage engine.<br />
<br />
regards<br />
Vengat]]></description>
            <dc:creator>vengadessa</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Tue, 07 Sep 2010 11:56:39 +0300</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,1256,1256#msg-1256</guid>
            <title>how to display information of last detected deadlock from   inodb SE in C or C++ (3 replies)</title>
            <link>http://forums.innodb.com/read.php?4,1256,1256#msg-1256</link>
            <description><![CDATA[ I am a newbie to this,so kindly help me out in writing a code for the following<br />
a C/C++ console application that connects to a MySQL server (with InnoDB plugin enabled) and prints information about the last detected deadlock (if any) by the InnoDB Storage Engine.<br />
eagerly awaiting for the response.<br />
<br />
<br />
<br />
<br />
kindly help me  with your suggestions to crack down this]]></description>
            <dc:creator>sumants88</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Tue, 03 Aug 2010 12:28:40 +0300</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,1255,1255#msg-1255</guid>
            <title>Why is this deadlocking? (2 replies)</title>
            <link>http://forums.innodb.com/read.php?4,1255,1255#msg-1255</link>
            <description><![CDATA[ Hi There,<br />
<br />
I am getting a deadlock on a SELECT statement that appears to be deadlocked by a DELETE statement.  Looking at the below statements in the log, shouldn't SELECT statements not be deadlocked by a parallel delete statement?  <br />
<br />
In scenario (SELECT content from a database while another process has an exclusive lock during a large DELETE), the way I read the log below is that the noted DELETE statement in transaction 2 (which is part of a series of DML that is wrapped in a transaction) has unknowingly put the table in an exclusive lock which deadlocks the SELECT statement.  If I am reading this right, I just don’t understand why this is the case as it seems to violate InnoDB’s MVCC support?  (transaction isolation set to “REPEATABLE-READ”).  <br />
<br />
I'm obviously misunderstanding something here.  What is the best way to defend against such deadlocks?  Rinse and repeat the SELECT statement?<br />
<br />
Thanks,<br />
<br />
Mike<br />
<br />
<br />
<pre class="bbcode">
------------------------
LATEST DETECTED DEADLOCK
------------------------
100722  9:11:30
*** (1) TRANSACTION:
TRANSACTION 0 79863245, ACTIVE 5 sec, process no 30967, OS thread id 1180342592 starting index read
mysql tables in use 4, locked 2
LOCK WAIT 6 lock struct(s), heap size 3024, 718 row lock(s)
MySQL thread id 582442, query id 30899169 172.30.5.38 abc preparing
SELECT
               m.id,
               m.default_alloy,
               m.description,
               m.screen_message,
               d.name as department_name,
               d.id as department_id,
               m.requires,
               m.optional,
               m.force_single_unit,
               m.turn_around_days,
               m.sort_rank
           FROM
               abc_lab_800020.macros as m
           LEFT JOIN
               abc_lab_800020.departments as d
           ON
               m.department_id = d.id WHERE
                           m.id IN (
                               SELECT DISTINCT
                                   p.macro_id
                               FROM
                                   abc_lab_800020.macros_practices as p
                               WHERE
                                   p.practice_id = '2353'
                           )
 
*** (1) WAITING FOR THIS LOCK TO BE GRANTED:
RECORD LOCKS space id 14132 page no 232 n bits 1120 index `IDX_macros_practices_1` of table `abc_lab_800020`.`macros_practices` trx id 0 79863245 lock mode S waiting
Record lock, heap no 2 PHYSICAL RECORD: n_fields 2; compact format; info bits 32
0: len 3; hex 414142; asc AAB;; 1: len 6; hex 000002b0a22c; asc      ,;;

*** (2) TRANSACTION:
TRANSACTION 0 79863205, ACTIVE 23 sec, process no 30967, OS thread id 1179810112 starting index read, thread declared inside InnoDB 500
mysql tables in use 1, locked 1
875 lock struct(s), heap size 161776, 191899 row lock(s), undo log entries 127795
MySQL thread id 582439, query id 30915183 172.30.5.38 abc updating
DELETE FROM
                           abc_lab_800020.macros_practices
                       WHERE
                           macro_id = 'A'
*** (2) HOLDS THE LOCK(S):
RECORD LOCKS space id 14132 page no 232 n bits 1152 index `IDX_macros_practices_1` of table `abc_lab_800020`.`macros_practices` trx id 0 79863205 lock_mode X
Record lock, heap no 1 PHYSICAL RECORD: n_fields 1; compact format; info bits 0
0: len 8; hex 73757072656d756d; asc supremum;;

....

</pre>]]></description>
            <dc:creator>michaelcaplan</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Fri, 27 Aug 2010 09:46:55 +0300</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,1250,1250#msg-1250</guid>
            <title>problem about redo log file write and flush (3 replies)</title>
            <link>http://forums.innodb.com/read.php?4,1250,1250#msg-1250</link>
            <description><![CDATA[ i am reading the code about innobase redo log,and i am confused when i see the code following<br />
here &quot;write&quot; means write from buffer pool to redo log buffer in memory ?<br />
and &quot;flush&quot; is transform the data from redo log buffer to redo log file? <br />
ut_dulint_cmp(log_sys-&gt;current_flush_lsn, lsn)  means i do not need call another write and flush opertion,because the lsn already being written?<br />
<br />
 if (log_sys-&gt;n_pending_writes &gt; 0) {<br />
                /* A write (+ possibly flush to disk) is running */<br />
<br />
                if (flush_to_disk<br />
                    &amp;&amp; ut_dulint_cmp(log_sys-&gt;current_flush_lsn, lsn)<br />
                    &gt;= 0) {<br />
                        /* The write + flush will write enough: wait for it to<br />
                        complete  */<br />
<br />
                        goto do_waits;<br />
                }<br />
<br />
                if (!flush_to_disk<br />
                    &amp;&amp; ut_dulint_cmp(log_sys-&gt;write_lsn, lsn) &gt;= 0) {<br />
                        /* The write will write enough: wait for it to<br />
                        complete  */<br />
<br />
                        goto do_waits;<br />
                }]]></description>
            <dc:creator>justlooks</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Sat, 28 Aug 2010 12:31:55 +0300</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,1245,1245#msg-1245</guid>
            <title>i wanna know that the process when innodb shutdown (no replies)</title>
            <link>http://forums.innodb.com/read.php?4,1245,1245#msg-1245</link>
            <description><![CDATA[ i just found that these params which affect innodb shutdown operation<br />
<br />
one is innodb_fast_shutdown ,the other is innodb_max_dirty_pages_pct<br />
<br />
but i wanna know the detail of innodb shutdown <br />
<br />
thx all~]]></description>
            <dc:creator>dehua.yeung</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Mon, 21 Jun 2010 09:27:38 +0300</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,1239,1239#msg-1239</guid>
            <title>innodb fk problem (1 reply)</title>
            <link>http://forums.innodb.com/read.php?4,1239,1239#msg-1239</link>
            <description><![CDATA[ Hello! <br />
<br />
I am new to mysql and I have some problems with my database. I am trying to connect tables created with InnoDB. The following example is't working and I can't fix it. <br />
<br />
[<a href="http://i46.tinypic.com/zxktoz.jpg" rel="nofollow" >i46.tinypic.com</a>] <br />
<br />
It seems that two FK from two different tables can't refer to a single PK. Do you have any suggestions?]]></description>
            <dc:creator>stoimenn</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Mon, 14 Jun 2010 09:15:46 +0300</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,1237,1237#msg-1237</guid>
            <title>Where is foreign key column name stored in system tables? (1 reply)</title>
            <link>http://forums.innodb.com/read.php?4,1237,1237#msg-1237</link>
            <description><![CDATA[ If a foreign key is defined on one of my tables what system table (INFORMATION_SCHEMA) would I query to see that foreign key column name?  I see a INFORMATION_SCHEMA table called referential_constraints that has the table name, foreign key name, and referenced table name but not the foreign key column name.<br />
<br />
I'm working off of InnoDB (it's the default storage engine set up on the MySQL server).<br />
<br />
What I want to do is programatically generate a list of multiple &quot;alter table add fk&quot; commands based on what currently exists in the database, then drop those fk's, the recreate them all based on that script I created.<br />
<br />
Thanks in advance,<br />
Wes]]></description>
            <dc:creator>wesandmarlene</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Tue, 08 Jun 2010 22:43:18 +0300</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,1228,1228#msg-1228</guid>
            <title>What is different between &quot;Innodb_pages_written&quot; and &quot;Innodb_data_written&quot; ? (2 replies)</title>
            <link>http://forums.innodb.com/read.php?4,1228,1228#msg-1228</link>
            <description><![CDATA[ We have two mysql server using innodb(built-in, not plugin)<br />
<br />
And, I gathered innodb global status value with (show global status like 'Innodb%') command.<br />
<br />
Here is the summary of two server ( Master  / Slave)<br />
<br />
----------------------------------------------------------------------------------------<br />
                                    Master            Slave            Master-Slave Diff<br />
                                    Diff/40Min        Diff/40Min        <br />
----------------------------------------------------------------------------------------<br />
Innodb_buffer_pool_write_requests   10556778          8336772          2220006<br />
Innodb_data_read                    73318400          54984704         18333696<br />
Innodb_data_reads                   4106              3164             942<br />
Innodb_data_writes                  1870504           958338           912166<br />
Innodb_data_written                 3107019776        2623567360       483452416<br />
Innodb_log_writes                   1820875           910543           910332<br />
Innodb_os_log_written               1550690304        898537472        652152832<br />
Innodb_pages_read                   4474              3356             1118<br />
Innodb_pages_written                94984             105281           -10297<br />
----------------------------------------------------------------------------------------<br />
<br />
And, I already read MySQL manaual, But I don't get it what they mean exactly.<br />
<br />
1. in &quot;Innodb_pages_written&quot;, what page mean ? <br />
   (in innodb_buffer_pool there are a lot of datas like undo, insert-buffer, cached-data, cached-index, ...)<br />
   does it mean all pages located in innodb buffer pool ?, <br />
   I think all those data(data, index, insert-buffer, undo) is managed with page(block) unit. is it right?<br />
<br />
2. what is different between &quot;Innodb_pages_written&quot; and &quot;Innodb_data_written&quot; ? <br />
<br />
3. How do I compare undo data's size written to disk or created in buffer pool ?<br />
   in &quot;show engine innodb status&quot;, I can't get absolute number against some time of two different machine.<br />
   above command show us the number of random period. <br />
<br />
4. Where can I detail explanation about the results of &quot;show global status like 'Innodb%'&quot; command ?.<br />
   Not Mysql manual, I think Innodb doesn't have enough manual about internal part.<br />
<br />
<br />
I'm not good at english. sorry for that.<br />
Really thanks.]]></description>
            <dc:creator>toto</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Thu, 03 Jun 2010 12:11:05 +0300</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,1223,1223#msg-1223</guid>
            <title>Error while creating the table in InnoDB (no replies)</title>
            <link>http://forums.innodb.com/read.php?4,1223,1223#msg-1223</link>
            <description><![CDATA[ Hi, <br />
<br />
I am getting the following error while creating the table with more than 70 foreign keys.can u please suggest how to create the table with more than 64 foreign keys. <br />
<br />
Is there any way to increase the No of keys?I am using the My SQL 5.5 version. <br />
I am migrating the database from Oracle to My SQL.one of my table in Oracle schema having the 100 foreign keys.I need to use the same table structure with 100 foreign keys.]]></description>
            <dc:creator>krishnareddyb</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Fri, 07 May 2010 09:19:56 +0300</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,1209,1209#msg-1209</guid>
            <title>And оne more thing. About row physical structure. (1 reply)</title>
            <link>http://forums.innodb.com/read.php?4,1209,1209#msg-1209</link>
            <description><![CDATA[ Hi!<br />
<br />
So,<br />
<br />
[<a href="http://dev.mysql.com/doc/refman/5.1/en/innodb-physical-record.html" rel="nofollow" >dev.mysql.com</a>]<br />
<br />
says that<br />
<br />
&quot;the variable-length part of the record header contains a bit vector for indicating NULL columns&quot;<br />
<br />
But in rows of the tables created by me this array is absent. Therefore I can't get size of the record's header and correct distribute row data between table columns. Thus, where in MySQL 5.1 row headers (or maybe page headers) is bit vector for indicating NULL columns?<br />
<br />
Sorry for my english,<br />
Thanks]]></description>
            <dc:creator>CopperbearD</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Wed, 28 Apr 2010 15:23:37 +0300</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,1208,1208#msg-1208</guid>
            <title>About pages &amp; tables (no replies)</title>
            <link>http://forums.innodb.com/read.php?4,1208,1208#msg-1208</link>
            <description><![CDATA[ Hi!<br />
Please, can you explain me how to compare page with table in MySQL 5.1 and higher? <br />
Situation: we have some &lt;tablename&gt; and file &quot;ibdata1&quot;. It is necessary to receive the list of the pages containing data of this table.<br />
Thanks.]]></description>
            <dc:creator>CopperbearD</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Mon, 26 Apr 2010 16:21:10 +0300</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,1195,1195#msg-1195</guid>
            <title>Mark a Clustered Node as deleted - without deleting the node (1 reply)</title>
            <link>http://forums.innodb.com/read.php?4,1195,1195#msg-1195</link>
            <description><![CDATA[ I have a very detailed question, perhaps anybody can answer it:<br />
<br />
Is it possible to delete a node (or mark as deleted) in the Clustered Index and to find this specific node afterwards again?<br />
<br />
I have the following problem: I've introduced a new command in InnoDB - so called &quot;Shredder&quot;. This command updates a field with random characters and deletes the entire row afterwards. Now I wan't to run again this Shredder command, but unfortunately this is not possible, since the index does not exist and the row couldn't be found. <br />
<br />
Do you have any ideas how I can find a deleted row or just to mark this row as deleted, but without deleting it in the Index?<br />
<br />
Gerhard, <a href="http://www.textilshop.at" rel="nofollow" >www.textilshop.at</a>]]></description>
            <dc:creator>gsoellradl</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Mon, 19 Apr 2010 08:58:00 +0300</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,1193,1193#msg-1193</guid>
            <title>shared tablespace (1 reply)</title>
            <link>http://forums.innodb.com/read.php?4,1193,1193#msg-1193</link>
            <description><![CDATA[ Innodb needs some process to periodically clear out the shared table space<br />
so the space can be reclaimed. This is the number one issue facing users today.<br />
The performance of this engine is great now improvements have been made.<br />
Please fix this issue. Other storage engines have no issue with this. <br />
Will this be fixed or do we need to switch to PrimeBase XT or Maria?]]></description>
            <dc:creator>lighttalk</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Fri, 16 Apr 2010 06:56:00 +0300</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,1182,1182#msg-1182</guid>
            <title>Queries with large row lock(s) values and no DB activity (no replies)</title>
            <link>http://forums.innodb.com/read.php?4,1182,1182#msg-1182</link>
            <description><![CDATA[ During benchmarking our system appears to hang for a period of time (e.g. 15-30 seconds).<br />
  <br />
SHOW ENGINE INNODB STATUS shows a number of transactions that have a number of row locks associated with them that appear to be the cause of contention however I am unable to determine from current statistics or settings any apparent reason for this occurance.<br />
 <br />
There are no apparent I/O bottlenecks (just a drop in disk I/O)<br />
<br />
An extract of INNODB STATUS shows three tables in active transactions.<br />
<br />
*  table1 has a number of DELETE statements. each is deleting on the first part of a two part PK, affecting ~130 rows for each delete<br />
* table2 has a number of multi value INSERTS, all active 9 seconds, one &quot;freeing items&quot; while one with a large number of row locks, and one appearing blocked in some way.<br />
* table3 has one INSERT also at 9 seconds?<br />
<br />
Any assistance in trying to understand more the cause and possible remedy greatly appreciated.<br />
<br />
---TRANSACTION 0 480082023, ACTIVE 1 sec, process no 16478, OS thread id 1207978336<br />
mysql tables in use 1, locked 1<br />
2 lock struct(s), heap size 368, 1 row lock(s)<br />
MySQL thread id 1136, query id 518011 xxx 10.120.229.37 db_2010 end<br />
DELETE FROM table1 WHERE TABLE_ID=2934<br />
---TRANSACTION 0 480082020, ACTIVE 1 sec, process no 16478, OS thread id 1206913376<br />
mysql tables in use 1, locked 1<br />
6 lock struct(s), heap size 1216, 139 row lock(s), undo log entries 135<br />
MySQL thread id 1132, query id 518007 xxx 10.120.229.37 db_2010 end<br />
DELETE FROM table1 WHERE TABLE_ID=2619<br />
---TRANSACTION 0 480082019, ACTIVE 1 sec, process no 16478, OS thread id 1207445856<br />
mysql tables in use 1, locked 1<br />
6 lock struct(s), heap size 1216, 139 row lock(s), undo log entries 135<br />
MySQL thread id 1134, query id 518006 xxx 10.120.229.37 db_2010 end<br />
DELETE FROM table1 WHERE TABLE_ID=2569<br />
---TRANSACTION 0 480082018, ACTIVE 1 sec, process no 16478, OS thread id 1206647136<br />
mysql tables in use 1, locked 1<br />
6 lock struct(s), heap size 1216, 139 row lock(s), undo log entries 135<br />
MySQL thread id 1131, query id 518005 xxx 10.120.229.37 db_2010 end<br />
DELETE FROM table1 WHERE TABLE_ID=2551<br />
---TRANSACTION 0 480082017, ACTIVE 1 sec, process no 16478, OS thread id 1195198816<br />
mysql tables in use 1, locked 1<br />
6 lock struct(s), heap size 1216, 139 row lock(s), undo log entries 135<br />
MySQL thread id 895, query id 518004 xxx 10.120.229.37 db_2010 end<br />
DELETE FROM table1 WHERE TABLE_ID=2586<br />
---TRANSACTION 0 480082016, ACTIVE 1 sec, process no 16478, OS thread id 1207179616<br />
mysql tables in use 1, locked 1<br />
5 lock struct(s), heap size 1216, 138 row lock(s), undo log entries 135<br />
MySQL thread id 1133, query id 518003 xxx 10.120.229.37 db_2010 end<br />
DELETE FROM table1 WHERE TABLE_ID=2539<br />
<br />
<br />
---TRANSACTION 0 480081823, ACTIVE 9 sec, process no 16478, OS thread id 1190140256<br />
mysql tables in use 1, locked <br />
11 lock struct(s), heap size 368, 0 row lock(s), undo log entries 420<br />
MySQL thread id 894, query id 517710 xxx 10.120.229.37 db_2010 update<br />
INSERT INTO table2 ...<br />
<br />
---TRANSACTION 0 480081822, ACTIVE 9 sec, process no 16478, OS thread id 1176295776<br />
mysql tables in use 1, locked 1<br />
5 lock struct(s), heap size 1216, 420 row lock(s), undo log entries 420<br />
MySQL thread id 334, query id 517708 xxx 10.120.229.37 db_2010 update<br />
INSERT INTO table2 ...<br />
<br />
---TRANSACTION 0 480081821, ACTIVE 9 sec, process no 16478, OS thread id 1197328736<br />
mysql tables in use 1, locked 1<br />
1 lock struct(s), heap size 368, 0 row lock(s), undo log entries 420<br />
MySQL thread id 925, query id 517704 xxx 10.120.229.37 db_2010 update<br />
INSERT INTO table3 ....<br />
<br />
---TRANSACTION 0 480081819, ACTIVE 9 sec, process no 16478, OS thread id 1169107296<br />
mysql tables in use 1, locked 1<br />
1 lock struct(s), heap size 368, 0 row lock(s), undo log entries 420<br />
MySQL thread id 329, query id 517699 xxx 10.120.229.37 db_2010 freeing items<br />
INSERT INTO table2 ...<br />
<br />
<br />
Environment<br />
<br />
MySQL 5.1.37<br />
RHEL 5 <br />
64bit<br />
innodb settings<br />
<br />
<br />
$ grep innodb my.cnf<br />
default-storage_engine = innodb<br />
innodb_buffer_pool_size = 8G<br />
innodb_additional_mem_pool_size = 20M<br />
innodb_log_buffer_size = 8M<br />
innodb_data_home_dir = /mysql/data<br />
innodb_data_file_path = ibdata01:10G;ibdata02:10G:autoextend<br />
innodb_log_group_home_dir = /mysql/iblogs/<br />
innodb_log_file_size = 512M<br />
innodb_log_files_in_group = 2<br />
innodb_flush_log_at_trx_commit = 2<br />
innodb_lock_wait_timeout = 20<br />
innodb_thread_concurrency = 0<br />
innodb_support_xa = 0]]></description>
            <dc:creator>ronaldbradford</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Thu, 08 Apr 2010 23:29:16 +0300</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,1162,1162#msg-1162</guid>
            <title>problem while running mysqld-max --console (no replies)</title>
            <link>http://forums.innodb.com/read.php?4,1162,1162#msg-1162</link>
            <description><![CDATA[ Hi when i am giving mysqld-max --console i am getting the error<br />
PLESE TREAT THIS AS MOST URGENT AS I AM NOT ABLE TO DO ANY THING SINCE 3 DAYS<br />
<br />
&quot;cannot initialise innoDB as 'innodb_data_file_path' is not set&quot;....<br />
<br />
<br />
i am using window 2003 server<br />
my.ini file is placed in c:/WINDOWS<br />
AND MY.INI IS AS SHOWN BELOW<br />
<br />
<br />
<br />
#This File was made using the WinMySQLAdmin 1.4 Tool<br />
#4/28/2002 6:19:56 PM<br />
<br />
#Uncomment or Add only the keys that you know how works.<br />
#Read the MySQL Manual for instructions<br />
<br />
<br />
[mysqld]<br />
basedir=C:/mysql<br />
bind-address=127.0.0.1<br />
datadir=C:/mysql/data<br />
port=3306<br />
<br />
<br />
set-variable=max_allowed_packet=15M<br />
default-table-type=innodb<br />
set-variable = max_connections=200<br />
set-variable = record_buffer=1M<br />
set-variable = sort_buffer=1M<br />
set-variable = key_buffer=30M<br />
set-variable = table_cache=40M<br />
<br />
innodb_data_home_dir=C:\mysql\ibdata<br />
innodb_data_file_path=ibdata1:512M;<br />
<br />
#use the following if the table type is innodb<br />
<br />
#buffer size should be 50-80% of the physical memory<br />
<br />
set-variable = innodb_buffer_pool_size=150M<br />
set-variable = innodb_additional_mem_pool_size=10M<br />
innodb_log_group_home_dir= C:\mysql\iblogs<br />
innodb_log_arch_dir=C:\mysql\iblogs<br />
innodb_log_archive=0<br />
set-variable = innodb_log_files_in_group=3<br />
<br />
#lof file size should be 15% of the buffer pool size<br />
<br />
set-variable = innodb_log_file_size=20M<br />
set-variable = innodb_log_buffer_size=8M<br />
innodb_flush_log_at_trx_commit=0<br />
<br />
set-variable = innodb_file_io_threads=14<br />
set-variable = innodb_lock_wait_timeout=50<br />
skip-bdb<br />
<br />
<br />
[WinMySQLadmin]<br />
Server=C:/mysql/bin/mysqld-max.exe<br />
user=root<br />
password=]]></description>
            <dc:creator>pagadalasatyam_84</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Wed, 24 Mar 2010 08:57:04 +0200</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,1142,1142#msg-1142</guid>
            <title>Deadlock caused by Foreign key insert S lock not escalating to update X (1 reply)</title>
            <link>http://forums.innodb.com/read.php?4,1142,1142#msg-1142</link>
            <description><![CDATA[ I am having trouble understanding why a particular deadlock situation occurs. The setup requires two tables:<br />
<br />
<pre class="bbcode">
CREATE TABLE parent (
       id INTEGER PRIMARY KEY,
       name varchar(10)
) ENGINE=InnoDB;

CREATE TABLE child (
       id INTEGER PRIMARY KEY,
       parent_id INTEGER NOT NULL,
       FOREIGN KEY (parent_id) REFERENCES parent(id)
) ENGINE=InnoDB;

insert into parent (id, name) values (1, 'charlie');</pre>
<br />
Creating the deadlock requires two threads, and can be done by opening two mysql command line sessions as follows:<br />
<br />
<pre class="bbcode">
-- in thread 1:
SET autocommit=0;
INSERT INTO child VALUES (1,1);
-- in thread 2:
SET autocommit=0;
UPDATE parent SET name ='alice' WHERE id = 1;</pre>
<br />
At this point, thread 2 blocks and is waiting for an exclusive row lock so it can update, and thread 1 is holding a shared row lock due to the foreign key on the insert. If we now do the following in thread 1, it will cause a deadlock to be detected, and for thread 2 to be rolled back.<br />
<br />
<pre class="bbcode">
UPDATE parent SET name ='bob' WHERE id = 1;</pre>
<br />
My question is, since thread 1 already holds a shared lock on the row why doesn't InnoDB just give it the exclusive lock? Why is it necessary to roll back thread 2? Couldn't thread 1 complete, and then thread 2 could be given the exclusive lock?]]></description>
            <dc:creator>Alok</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Sun, 14 Mar 2010 07:52:11 +0200</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,1138,1138#msg-1138</guid>
            <title>innodb wont start in mysql 5.1 (1 reply)</title>
            <link>http://forums.innodb.com/read.php?4,1138,1138#msg-1138</link>
            <description><![CDATA[ hi<br />
i have prob with innodb under ubuntu 9.10 plateform, afetr install ubuntu 9.10 in a proliant server i ve tryed to install mysql 5.1 but when started, mysql show this error message <br />
<pre class="bbcode">
ParamÃ©trage de mysql-server-5.1 (5.1.37-1ubuntu5.1) ...
 * Stopping MySQL database server mysqld                                 [ OK ]
100309 18:37:33 [Note] Plugin 'FEDERATED' is disabled.
100309 18:37:33  InnoDB: Started; log sequence number 0 44233
100309 18:37:33  InnoDB: Starting shutdown...
100309 18:37:35  InnoDB: Shutdown completed; log sequence number 0 44233
100309 18:37:35 [Warning] Forcing shutdown of 1 plugins
 * Starting MySQL database server mysqld</pre>
<br />
i need innodb as default storage engine to restore my db<br />
<br />
thks to help me]]></description>
            <dc:creator>khaled_jamel</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Tue, 16 Mar 2010 12:54:46 +0200</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,1111,1111#msg-1111</guid>
            <title>Innodb scalability related to autoinc (2 replies)</title>
            <link>http://forums.innodb.com/read.php?4,1111,1111#msg-1111</link>
            <description><![CDATA[ Hello,<br />
<br />
I am using Innodb engine for all tables, and uses &quot;unsigned big int&quot; as primary key with auto increment for these tables with MySql version 5.0.77.  One of such table with high concurrency [ 4-100 concurrent threads ]  is showing some issues under load , also this table does not have secondary index and primarily we just insert record into this table. The table is around 60GB and have more than 1/2 billion records. The insert speeds varies from 40 - 500 Inserts/seconds depending on the traffic surge. Recently we are seeing lot of these [80+] inserts blocked on following states , as per &quot;show innodb engine status \G&quot; report <br />
============<br />
trx id 3 3049405756 lock mode AUTO-INC waiting<br />
OS thread id 1186077008 setting auto-inc lock<br />
trx id 3 3049405755 lock mode AUTO-INC waiting<br />
...<br />
....<br />
OS thread id 1163397456 setting auto-inc lock<br />
=======<br />
 I know that Innodb uses table lock for autoincrement values, and like to know<br />
  1] Is there any hard limit under which the table lock could be severe problem ? If so let us know.<br />
  2] In order to eliminate the table lock because of autoinc, is there any other alternatives for primary key which would result in very high insert but yet no performance penalty as well as table fragmentation ? <br />
  3] Any other recommendation like going to 5.1 where the problem is solved or sharding ?<br />
<br />
Thanks for your help !<br />
Ramaki]]></description>
            <dc:creator>ramaki</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Mon, 22 Feb 2010 19:02:07 +0200</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,1107,1107#msg-1107</guid>
            <title>InnoDB Lock Manager Design Question (no replies)</title>
            <link>http://forums.innodb.com/read.php?4,1107,1107#msg-1107</link>
            <description><![CDATA[ Hello InnoDB folks.  I've been studying InnoDB's source code as part<br />
of graduate school research project, and I was hoping that one of you<br />
could answer a design question.  I'm asking this as a matter of<br />
personal curiosity.<br />
<br />
I've found a lot of very clever optimizations in InnoDB's source code.<br />
For example, the bitmap structure for locks, the way that mem_heap<br />
tries to minimize allocation by letting callers provide an initial<br />
chunk of memory, and the linked list &quot;heap&quot; for page tuples.  They're<br />
all pretty neat ideas.<br />
<br />
My question has to do with InnoDB's kernel_mutex, the global mutex<br />
that protects the lock manager.  I would have expected to see one<br />
mutex per hash table cell (a-la Gray &amp; Reuter pg. 471), rather than<br />
one mutex for the whole lock manager.  Given the amount of attention<br />
to detail in InnoDB's source code, I'm sure this was a very<br />
intentional decision.<br />
<br />
I also noticed a line that's commented out in lock_sys_create()<br />
<br />
   /* hash_create_mutexes(lock_sys-&gt;rec_hash, 2, SYNC_REC_LOCK); */<br />
<br />
This seems to imply that at one point, someone considered the idea of<br />
one mutex per lock manager hash cell, then changed their mind.<br />
<br />
So I was wondering -- why did you choose the single mutex option?  Did<br />
that perform better than per-cell mutexes?  Were per-cell mutexes too<br />
error prone to work with?<br />
<br />
Thanks in advance for your time.<br />
<br />
Steve Revilak<br />
UMass Boston]]></description>
            <dc:creator>srevilak</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Tue, 16 Feb 2010 03:14:36 +0200</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,1084,1084#msg-1084</guid>
            <title>innodb and sorting (no replies)</title>
            <link>http://forums.innodb.com/read.php?4,1084,1084#msg-1084</link>
            <description><![CDATA[ Hi,<br />
I'm making a website which has a ranked ladder. Players will have their player id and be given an ELO rating. There will be frequent adjustments to the ELO rating and consequently their rank. I'm a little new with MySQL and so my query looks like SELECT * FROM players ORDER BY rating DESC LIMIT 100, 50. <br />
The user base would probably be in the thousands range with probably a few hundred active at any given time.<br />
<br />
I'm wondering if the sorting would be excessive given that the rankings would change often? Would it make a difference to use Innodb or MyISAM?]]></description>
            <dc:creator>hath</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Fri, 12 Feb 2010 04:43:01 +0200</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,1044,1044#msg-1044</guid>
            <title>Query cost estimation (no replies)</title>
            <link>http://forums.innodb.com/read.php?4,1044,1044#msg-1044</link>
            <description><![CDATA[ I created 1 GB TPC-H instance on both Windows and Linux with the same indexes. In both instances the storage engine is selected as InnoDB.<br />
When I examine the query plans and cost estimations, I notice that the cost estimations on Linux is much more accurate than the one on windows. And running the query on windows takes much much longer than the one on linux.<br />
The version of mysql on windows is 5.0.24 and 5.1.38 on linux.<br />
I wonder if there has been significant changes in the query engine between these versions? or something else is goin on and I dont know what that is.. (I think I should try with the same versions but I dont want to spend some more time for recompilation and installation steps)<br />
any idea or suggestion?]]></description>
            <dc:creator>Richard</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Sat, 06 Feb 2010 20:53:27 +0200</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,1041,1041#msg-1041</guid>
            <title>Extent size in innodb (no replies)</title>
            <link>http://forums.innodb.com/read.php?4,1041,1041#msg-1041</link>
            <description><![CDATA[ Is it possible to change the default extent size in InnoDB? Either internally (from the source code) or externally? <br />
and anybody knows the extent size in MyIsam tables? <br />
Thanks.]]></description>
            <dc:creator>Richard</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Fri, 05 Feb 2010 23:10:01 +0200</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,1021,1021#msg-1021</guid>
            <title>Can't open file (7 replies)</title>
            <link>http://forums.innodb.com/read.php?4,1021,1021#msg-1021</link>
            <description><![CDATA[ Using MySql 4.1.7-nt.<br />
<br />
Getting this errors:<br />
<br />
100123  8:10:18 [ERROR] C:\Program Files\MySQL\MySQL Server 4.1\bin\mysqld-nt: Can't open file: 'custcodes_suppliers.InnoDB' (errno: 1)<br />
100123  8:10:22  InnoDB error:<br />
Cannot find table mthosp/custcodes_suppliers from the internal data dictionary<br />
of InnoDB though the .frm file for the table exists. Maybe you<br />
have deleted and recreated InnoDB data files but have forgotten<br />
to delete the corresponding .frm files of InnoDB tables, or you<br />
have moved .frm files to another database?<br />
Look from section 15.1 of [<a href="http://www.innodb.com/ibman.html" rel="nofollow" >www.innodb.com</a>]<br />
how you can resolve the problem.<br />
<br />
Please help, can not delete / alter the table or do abackup always fail on custcodes_suppliers table.<br />
<br />
Can not find <br />
section 15.1 of [<a href="http://www.innodb.com/ibman.html" rel="nofollow" >www.innodb.com</a>]<br />
<br />
PLEASE PLEASE HELP!]]></description>
            <dc:creator>Tal Ben-Gal</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Tue, 31 Aug 2010 05:40:38 +0300</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,1013,1013#msg-1013</guid>
            <title>InnoDB engine not available when installing MySQL silently (1 reply)</title>
            <link>http://forums.innodb.com/read.php?4,1013,1013#msg-1013</link>
            <description><![CDATA[ After quietly instlling 64-bit version of MySQL with the following command, the InnoDB engine was unavailable.<br />
command: msiexec /i mysql-essential-5.0.83-winx64.msi /qn<br />
<br />
When running MySQLInstanceConfig, all the InnoDB options are greyed-out suggesting InnoDB is not available. I did not face this problem when installing MySQL using the GUI with all the default options.<br />
<br />
I was using MySQL-5.0.83 on Windows Server 2003.<br />
<br />
Further to this, are there InnoDB-related command-line parameters that can be passed to the msi file?<br />
<br />
Thanks,<br />
Prashanta]]></description>
            <dc:creator>Prashanta</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Mon, 18 Jan 2010 10:18:06 +0200</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,998,998#msg-998</guid>
            <title>Does flipping autocommit have a performance hit? (3 replies)</title>
            <link>http://forums.innodb.com/read.php?4,998,998#msg-998</link>
            <description><![CDATA[ Hi<br />
<br />
I have a database that is performing a lot of &quot;set&quot; commands (almost half of incoming traffic).  From mysqlreport:<br />
<br />
__ Questions ___________________________________________________________<br />
Total           1.67G   706.9/s<br />
  <b>Com_        747.52M   317.2/s  %Total:  44.87</b><br />
  DMS         462.82M   196.4/s           27.78<br />
  QC Hits     429.17M   182.1/s           25.76<br />
  +Unknown     17.30M     7.3/s            1.04<br />
  COM_QUIT      9.15M     3.9/s            0.55<br />
snip<br />
Com_          747.52M   317.2/s           44.87<br />
  <b>set_option  633.80M   268.9/s           38.04</b><br />
  commit       35.57M    15.1/s            2.13<br />
  show_variab  34.50M    14.6/s            2.07<br />
<br />
The developers have told me their code wraps each block of commands from their application, even if a block consists of a single select, in:<br />
<br />
set autocommit=0;<br />
some statement(s);<br />
commit;<br />
set autocommit=1;<br />
<br />
So my question is does this frequent autocommit mode flipping have a performance hit on the database?  If yes, how might this manifest?  A brief explanation of what innodb does when asked to switch autocommit modes might help too :)<br />
<br />
Thanks in advance<br />
Justin.]]></description>
            <dc:creator>justinbennett</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Fri, 29 Jan 2010 06:19:25 +0200</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,995,995#msg-995</guid>
            <title>unable to lock ibdata1 (6 replies)</title>
            <link>http://forums.innodb.com/read.php?4,995,995#msg-995</link>
            <description><![CDATA[ Bonjour,<br />
<br />
I have this message :<br />
Unable to lock ./ibdata1<br />
<br />
more info :<br />
System Linux www2 2.6.18-6-amd64<br />
PHP Version 5.2.0-8+etch16<br />
Mysql 5.0.32-Debian_7etch11-log<br />
<br />
lsof /var/lib/mysql/ibdata1<br />
COMMAND PID USER FD TYPE DEVICE SIZE NODE NAME<br />
mysqld 30261 mysql 5uW REG 254,7 18874368 2459 /var/lib/mysql/ibdata1<br />
<br />
/etc/fstab :<br />
/dev/mapper/vg1-var2 /var reiserfs <br />
<br />
I read somewhere there is a bug with NFS partition, what about logical volume ?<br />
<br />
Thanks for your time,<br />
N]]></description>
            <dc:creator>nine</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Wed, 13 Jan 2010 16:42:18 +0200</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,993,993#msg-993</guid>
            <title>blocking query shows null at all (6 replies)</title>
            <link>http://forums.innodb.com/read.php?4,993,993#msg-993</link>
            <description><![CDATA[ I made some blocking happen, but the blocking query shows null at all from the below query.<br />
<br />
SELECT r.trx_id waiting_trx_id,  r.trx_mysql_thread_id waiting_thread,<br />
       r.trx_query waiting_query,<br />
       b.trx_id blocking_trx_id, b.trx_mysql_thread_id blocking_thread,<br />
       b.trx_query blocking_query<br />
FROM       information_schema.innodb_lock_waits w<br />
INNER JOIN information_schema.innodb_trx b  ON  b.trx_id = w.blocking_trx_id<br />
INNER JOIN information_schema.innodb_trx r  ON  r.trx_id = w.requesting_trx_id;]]></description>
            <dc:creator>larryloi</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Fri, 29 Jan 2010 13:07:46 +0200</pubDate>
        </item>
        <item>
            <guid>http://forums.innodb.com/read.php?4,987,987#msg-987</guid>
            <title>shared tablespace growth (5 replies)</title>
            <link>http://forums.innodb.com/read.php?4,987,987#msg-987</link>
            <description><![CDATA[ hi,<br />
<br />
i have been trying to understand why/when the shared tablespace is used, mostly to understand why it grows.<br />
<br />
i have a 500GB db, with a single table taking up 360GB, another couple taking up 50GB and then a bunch of others.<br />
<br />
i have innodb_file_per_table set, so each table has its own .ibd outside the shared tablespace.<br />
<br />
i have noticed if i do &quot;insert into foo select * from my_bigtable where ...&quot; if a large chunk of data is being inserted, the shared tablespace (i have a single ibdata1 file) suddenly grows.<br />
<br />
i try to avoid this by doing inserts in smaller chunks.<br />
<br />
recently i did a bunch of deletes (basically deleting data for some of our customers from the fact tables). i employed the same technique by deleting day by day (i.e. issuing delete queries in a loop 1 per day).<br />
<br />
however, i notice the shared tablespace file has grown again (from 40GB to 163GB in the 7 hours while i was running the deletes).<br />
<br />
the only thing that i have read is<br />
(http://dev.mysql.com/doc/refman/5.1/en/multiple-tablespaces.html) <br />
<br />
&quot;&quot;&quot; <br />
InnoDB always needs the shared tablespace because it puts its internal data dictionary and undo logs there. The .ibd files are not sufficient for InnoDB to operate. <br />
&quot;&quot;&quot; <br />
<br />
but i have also seen other seemingly contradictory statements... is this growth caused by undo logs? each of my delete statements are in a single transaction, so the undo logs should be purgable after each statement.<br />
<br />
below are some excerpts from my.cnf:<br />
&quot;&quot;&quot;<br />
innodb_buffer_pool_size         = 10240M<br />
innodb_additional_mem_pool_size = 256M<br />
# Set .._log_file_size to 25 % of buffer pool size<br />
innodb_log_file_size            = 1024M<br />
innodb_log_buffer_size          = 64M<br />
innodb_flush_log_at_trx_commit  = 2<br />
innodb_lock_wait_timeout        = 120<br />
<br />
innodb_flush_method             = O_DIRECT<br />
innodb_support_xa               = OFF<br />
innodb_doublewrite              = 0<br />
innodb_file_per_table           = 1<br />
innodb_log_files_in_group       = 2<br />
#Allow tmp tables in memory to grow up to 512*1024*1024 bytes before<br />
#they are converted into tmp tables on disk which seems to be expensive<br />
max_heap_table_size = 536870912<br />
tmp_table_size = 536870912<br />
&quot;&quot;&quot;<br />
<br />
would really appreciate an explanation or a pointer to the explanation (short of &quot;look in the source&quot;).<br />
<br />
thanks]]></description>
            <dc:creator>nishant.deshpande</dc:creator>
            <category>InnoDB Storage Engine</category>
            <pubDate>Fri, 15 Jan 2010 20:35:14 +0200</pubDate>
        </item>
    </channel>
</rss>
