3 messages in com.mysql.lists.mysqlInnobase table check.
FromSent OnAttachments
Peter Zaitsev15 Mar 2001 01:40 
Michael Widenius15 Mar 2001 05:54 
Heikki Tuuri15 Mar 2001 07:29 
Subject:Innobase table check.
From:Michael Widenius (mon@mysql.com)
Date:03/15/2001 05:54:08 AM
List:com.mysql.lists.mysql

Hi!

"Peter" == Peter Zaitsev <pz@spylog.ru> writes:

Peter> Hello Heikki, Peter> Finally I was able to check the innobase tables included into mysql Peter> 3.23.34. Well first several times I was quite happy about them, but Peter> later understud that this is because option --create-options is Peter> broken :)

Peter> So now I must say On my system innobase seems to work as bad as it Peter> worked before :(

Peter> - ATIS test fails.

<cut>

This is strange. I have run all benchmark tests with Innobase without any problems.

How did you compile MySQL and Innobase ? Can you send us the output from mysqlbug ?

Peter> - mysqld is restarted during alter table test and one more time Peter> during the tests:

Peter> Innobase: Started Peter> /usr/local/mysql/libexec/mysqld: ready for connections Peter> Innobase: Warning: out of memory in additional memory pool. Peter> Innobase: Innobase will start allocating memory from the OS. Peter> Innobase: You should restart the database with a bigger value in Peter> Innobase: the MySQL .cnf file for innobase_additional_mem_pool_size. Peter> 010314 19:00:00 Warning: Checking table: './oldgoodcounter/stop_recs' Peter> 010314 19:00:01 Warning: Checking table:
'./oldgoodcounter/registrants_stats'

<cut>

Peter> - It seems like error message for error 139 should be changed because Peter> it says about 16M there innobase and gemini has their own limits Peter> about it. Peter> 139 = Too big row (>= 16 M)

I shall fix that.

Peter> Now few words about reasons why this may happen (I'll try to check Peter> them out soon)

Peter> 1) I'm using 2.4.2 kernel,SMP - so there may be some incompatibilities. Peter> 2) I'm usung patched for 2GB limit GLIBC Peter> 3) I'm using ReiserFS file system. Peter> 4) The parameters I'm using. (Like bdb does not work with big Peter> tablecache)

It could be the parameters; I don't know about that.

What startup parameters are you using?

Peter> Anyway MYSQL with MYISAM works with no problem on this system, and I Peter> use the same system for production on 20 machines so this looks for me Peter> more like incomportability problem.

Regards, Monty