Re:InnoDB

2001-10-05 Thread Heikki Tuuri
Riccardi, your innodb_log_arch_dir is C:\InnoDB\iblogs but innodb_log_group_home_dir is C:\iblogs These should be the same. Your MySQL error log should contain an error message of this: C:\trybuild\client_debug>mysqld --console InnoDB: Error: you must set the log group home dir in my.cnf the

Re:InnoDB Recovery Job %

2001-09-24 Thread Heikki Tuuri
Alex, at the maximum it has to scan the whole size of its log files. In the 'applying log records to the database' phase it may take a long time because of disk i/o. Thus, it is really not possible to know the remaining time precisely since the amount of disk i/o may vary in different phases of

Re:InnoDB vs MyISAM on COUNT(*) ... WHERE ...

2001-09-18 Thread Heikki Tuuri
Sander, >I'm thinking of switching to InnoDB, however - my application does a few >COUNT(*) WHERE queries on large tables (somewhere between 50K >and 2M rows) > >I've read up on InnoDB and its issues with COUNT(*) on entire tables, >but is there a reason to assume that InnoDB is also sl

Re:Innodb restart problem

2001-08-13 Thread Heikki Tuuri
Xavier, I assume you have increased innodb_buffer_pool_size to 16M or bigger? The fsync problem fixed in .40b could cause that InnoDB was not able to flush modified pages quickly enough to disk, causing the error message 'Difficult to find free block..." Please download the brand-new 3.23.41

Re:InnoDB and table optimization?

2001-08-09 Thread Heikki Tuuri
Philip, InnoDB does not give an exact number of rows to the query analyzer. A problem is that in a multiversioned database different transactions will see a different number of rows in the table, and a single count of rows cannot be kept. It is an eternal problem in query optimization. Errors in

Re:innodb and raw disks

2001-07-17 Thread Heikki Tuuri
Hi! You should make an unformatted partition whose size is divisible by 1 MB. If fseek does not understand the partition size, we can disable the size check from InnoDB code. Please report if you succeed. If not, I change InnoDB-3.23.40 so that you can disable the file size check. I haver not t

Re:innodb crash

2001-07-04 Thread Heikki Tuuri
Julien, sorry that I did not notice your message earlier, but I was running tests on new big BLOBs, and a stress test passed now without errors :). But the problem you have is probably the rollback bug in .38. It has been fixed in .39. Upgrade to to .39 and try to start the database again. Rega

Re:InnoDB not enabled

2001-06-24 Thread Heikki Tuuri
Tomek, have you defined the InnoDB startup options in your my.cnf or my.ini? You can look at the manual at http://www.innodb.com about example settings for InnoDB in Windows. Did you compile MySQL yourself or are you using one of the binaries in the binary distribution? MySQL -Max 3.23.39 is alr

Re:innodb text fields

2001-06-14 Thread Heikki Tuuri
Ryan, I hope it pays to wait. If you compile MySQL/InnoDB yourself, you will get unlimited BLOB and TEXT fields within 16 days. I am writing the support for them now and I will release the source code at my website. It will also be in the next binary release of MySQL/InnoDB, unless Monty relea

Re:innodb question

2001-05-24 Thread Heikki Tuuri
Sommai, I have developed InnoDB on a computer with 32 MB memory and 1.2 GB disk. You can run it on 64 MB, but be careful when you specify the buffer pool, key cache and other memory pool parameters in my.cnf, so that you do not run out of physical memory. If Linux has to swap to disk, it can be v

Re:innodb bug

2001-05-23 Thread Miguel Angel Solórzano
At 18:10 23/5/2001 +0300, Heikki Tuuri wrote: Hi, >Claudio, > >in the source distribution the macro > >__NT__ > >is defined. That makes it work only on NT and 2000. >You should compile without __NT__ . See a posting by Miguel >about 5 days ago. If you don't have the VC++ compiler use mysqld-max.e

Re:innodb bug

2001-05-23 Thread Heikki Tuuri
Claudio, in the source distribution the macro __NT__ is defined. That makes it work only on NT and 2000. You should compile without __NT__ . See a posting by Miguel about 5 days ago. Regards, Heikki >Downloaded latest version of MySQL for Win95 (3.23.38-win) on May 21st. >After "mysqld" comm

Re:InnoDB problem

2001-05-16 Thread Heikki Tuuri
Jamie, I do not think it is the Reiser file system. Could you show what parameters you have in my.cnf and what is the size of the file ibdata1 (when looked with ls -l)? Regards, Heikki http://www.innobase.fi >I complied MySQL as a max binary. However when I gave all the variables for >InnoDB i

Re:InnoDB tablespace tools

2001-05-15 Thread Heikki Tuuri
Steve, I cannot promise any such tools in the next 6 months. My TODO list has the two important items currently: support for arbitrary size BLOBs and CHECK TABLE. The right way to proceed now is to test existing tools and their speed. Could you measure the speed of the following options in dump

Re:InnoDB and concurrent operations

2001-05-14 Thread Heikki Tuuri
Steve, on what operating system you are running? Was the rollback (and the table load) disk bound? Did you configure much memory to the InnoDB buffer pool? Was MySQL responsive during the table loading? Is MySQL responsive if you do a big rollback without cancelling (ctrl-c) the client process