Comments on: Upgrading from MySQL 4 – user stories http://blog.adishatz.net/?p=19 random thoughts about the LAMP platform Sat, 26 Jun 2010 15:21:14 +0000 http://wordpress.org/?v=2.9.2 hourly 1 By: Guillaume http://blog.adishatz.net/?p=19&cpage=1#comment-105 Guillaume Sat, 26 Jun 2010 15:21:14 +0000 http://blog.adishatz.net/?p=19#comment-105 Cindy, it means that your tables are probably still in the older 4.1 format. Retrocompatibility is a possibility, but I would recommend that you run the mysql_upgrade script to alter the tables to the new format. DECIMAL and VARCHAR fields have different implementations starting from 5.0. Please read http://dev.mysql.com/doc/refman/5.0/en/upgrading-from-previous-series.html for more information. Cindy,

it means that your tables are probably still in the older 4.1 format. Retrocompatibility is a possibility, but I would recommend that you run the mysql_upgrade script to alter the tables to the new format. DECIMAL and VARCHAR fields have different implementations starting from 5.0. Please read http://dev.mysql.com/doc/refman/5.0/en/upgrading-from-previous-series.html for more information.

]]> By: Cindy http://blog.adishatz.net/?p=19&cpage=1#comment-104 Cindy Fri, 25 Jun 2010 18:34:07 +0000 http://blog.adishatz.net/?p=19#comment-104 You mentioned, "Note that binary upgrades are still possible between 5.0 and 5.1 since the storage definitions haven’t changed." Are 4.1 and 5.5 storage definitions different? We've swaped in 5.5 MySQL binaries with a 4.1 datadir and haven't had any issues (only testing so far) You mentioned,

“Note that binary upgrades are still possible between 5.0 and 5.1 since the storage definitions haven’t changed.”

Are 4.1 and 5.5 storage definitions different? We’ve swaped in 5.5 MySQL binaries with a 4.1 datadir and haven’t had any issues (only testing so far)

]]> By: Guillaume http://blog.adishatz.net/?p=19&cpage=1#comment-45 Guillaume Mon, 11 Jan 2010 16:20:04 +0000 http://blog.adishatz.net/?p=19#comment-45 Hi Robert, thanks, that will be definitely helpful. I can't really say how much 4.x installations are still out and ready to migrate. There must be a lot of them remaining though ! Hi Robert,
thanks, that will be definitely helpful.
I can’t really say how much 4.x installations are still out and ready to migrate. There must be a lot of them remaining though !

]]> By: Robert Hodges http://blog.adishatz.net/?p=19&cpage=1#comment-39 Robert Hodges Fri, 08 Jan 2010 21:15:24 +0000 http://blog.adishatz.net/?p=19#comment-39 Thanks for the response. I just filed a JIRA (http://forge.continuent.org/jira/browse/TREP-394) to start replication from a specific location and can more or less guarantee it will get done in the next 30 days or so. We are looking hard at features to support complex topologies (e.g., SaaS data architectures) and this is one of the items that was already on my mental to-do list. On down-version replication--it more or less works now though we don't certify it very regularly. Do you have a sense of how widely 4.0/4.0 to 5.1 migrations are taking place? To be frank the v3.0 binlog support probably would not be too hard. Thanks for the response. I just filed a JIRA (http://forge.continuent.org/jira/browse/TREP-394)
to start replication from a specific location and can more or less guarantee it will get done in the next 30 days or so. We are looking hard at features to support complex topologies (e.g., SaaS data architectures) and this is one of the items that was already on my mental to-do list.

On down-version replication–it more or less works now though we don’t certify it very regularly. Do you have a sense of how widely 4.0/4.0 to 5.1 migrations are taking place? To be frank the v3.0 binlog support probably would not be too hard.

]]> By: Tech.Bitez » Upgrading MySQL 4 http://blog.adishatz.net/?p=19&cpage=1#comment-37 Tech.Bitez » Upgrading MySQL 4 Fri, 08 Jan 2010 18:00:28 +0000 http://blog.adishatz.net/?p=19#comment-37 [...] Today I found a couple of useful articles about upgrading MySQL 4. One is written by Peter Zaitsev and other by Guillaume Lefranc. [...] [...] Today I found a couple of useful articles about upgrading MySQL 4. One is written by Peter Zaitsev and other by Guillaume Lefranc. [...]

]]> By: Guillaume http://blog.adishatz.net/?p=19&cpage=1#comment-35 Guillaume Fri, 08 Jan 2010 13:43:11 +0000 http://blog.adishatz.net/?p=19#comment-35 Hi Robert, my point was to be able to start replicating from a specific binlog position from the master (like you would do with traditional mysql replication if you would start from an existing setup). The "down-version" replication feature is very interesting so let me know if you have any progress. Hi Robert,
my point was to be able to start replicating from a specific binlog position from the master (like you would do with traditional mysql replication if you would start from an existing setup). The “down-version” replication feature is very interesting so let me know if you have any progress.

]]> By: Robert Hodges http://blog.adishatz.net/?p=19&cpage=1#comment-33 Robert Hodges Fri, 08 Jan 2010 06:09:12 +0000 http://blog.adishatz.net/?p=19#comment-33 Hi! First of all thanks for the pointer to Tungsten. (It's my baby.) By the CHANGE MASTER comment, do you mean the ability to restart Tungsten from an arbitrary point in the slave? If not, could you elaborate? The real blocker though for Tungsten is that we don't support 4.0/4.1 logs yet. I'm kind of in denial on this but it sounds as if this case may be more common than we thought. We can replicate from 5.1/5.0 down to 4.x but not the other way around. Hi! First of all thanks for the pointer to Tungsten. (It’s my baby.) By the CHANGE MASTER comment, do you mean the ability to restart Tungsten from an arbitrary point in the slave? If not, could you elaborate?

The real blocker though for Tungsten is that we don’t support 4.0/4.1 logs yet. I’m kind of in denial on this but it sounds as if this case may be more common than we thought. We can replicate from 5.1/5.0 down to 4.x but not the other way around.

]]> By: Sheeri K. Cabral (Pythian) http://blog.adishatz.net/?p=19&cpage=1#comment-31 Sheeri K. Cabral (Pythian) Thu, 07 Jan 2010 18:26:51 +0000 http://blog.adishatz.net/?p=19#comment-31 The replication discrepancy between 4.1 and 5.1 is noted on the same page that it says MySQL supports replication between older masters and newer slaves: http://dev.mysql.com/doc/refman/5.1/en/replication-compatibility.html Also, I've found that comma joins are the cause of the most problems when upgrading. The VARCHAR and DECIMAL points are *extremely important*, but I've found that it is more difficult to fix (and find) the comma join problem. See: http://pento.net/2009/04/03/join-and-comma-precedence/ http://dev.mysql.com/doc/refman/5.0/en/join.html (search for "comma join" on the page) (and if you want a good rant on other reasons not to use comma joins, http://www.pythian.com/news/tag/comma-join/ ) The replication discrepancy between 4.1 and 5.1 is noted on the same page that it says MySQL supports replication between older masters and newer slaves:

http://dev.mysql.com/doc/refman/5.1/en/replication-compatibility.html

Also, I’ve found that comma joins are the cause of the most problems when upgrading. The VARCHAR and DECIMAL points are *extremely important*, but I’ve found that it is more difficult to fix (and find) the comma join problem.

See:
http://pento.net/2009/04/03/join-and-comma-precedence/
http://dev.mysql.com/doc/refman/5.0/en/join.html
(search for “comma join” on the page)

(and if you want a good rant on other reasons not to use comma joins, http://www.pythian.com/news/tag/comma-join/ )

]]>