Thread

Hi!

I’ve lost the reference (sorry), but I’ve learned from this forum that a 1.6.x to 2.4.0 database migration with non-UTF-8 character sets requires special attention by someone at Ellislab.

Is this correct? If so: at which point do you need the database dump?

My specific upgrade path has to be: 1.6.x—> 2.2.2—> 2.4.0.

Thanks a bunch… Nico

(edit: fix typo)

Hi Nico,

I don’t know specifically, but I will take this up to our devs for clarification.

My specific upgrade path has to be: 1.6.x—> 2.2.2—> 2.4.0.

Can I ask why you need to go to 2.2.2 first?

Hold tight while I get the devs on this.

Cheers,

Hi Dan,

Thanks for your response!

Can I ask why you need to go to 2.2.2 first?

Upgrading to 2.4.0 bombs out on a large database. Running the update.sh script succesfully and then refreshing the page in the browser… nothing happens: I get the same message.

Via 2.2.2 it works. 😊

Thanks… Nico

Hi Nico,

I got some clarification on your question.

Basically, the best plan here is to get your site into a development or local environment and run through the process. Take note of anything along the way that doesn’t work and come back to us with any errors.

Upgrading to 2.4.0 bombs out on a large database. Running the update.sh script succesfully and then refreshing the page in the browser… nothing happens: I get the same message.

How big is your database?

Cheers,

Hi Dan,

Sorry it took me a while to get back to you. Been crazy busy.

Basically, the best plan here is to get your site into a development or local environment and run through the process. Take note of anything along the way that doesn’t work and come back to us with any errors.

Somehow, over the last week, the ‘source’ database has shrunk in size from 320Mb to 89Mb.

1.6.6—> 2.4.0 is now a single step with this new, smaller database. That’s a big plus!

Now, all that remains, is the conversion of Polish and Russian texts.

Polish texts are mostly okay, but Russian texts are all question marks.

On 1.6.6, exp_weblog_data collation is latin1_swedish_ci. On 2.4.0, exp_channel_data collation is utf8_general_ci. Does that give you any clues?

Thank you so far!

Grtz… Nico

OK Nico,

I spoke with one of the devs, and the thinking is that since you are in the high ASCII char set already, that the updated is converting your UTF-8 encoded characters over again. There is a workaround, but I must stress that you try this on a test environment first.

You will need to start at the beginning with 1.6.6 again. Once you have that up and running, before you run the update to 2.4, you will need to edit a file.

In your copy of 2.4, go to /system/expressionengine/installer/updates/ud_200.php
Find:

foreach ($tables as $table)
  {
   $progress = "Converting Database Table {$table}: %s";
   $count  = $this->EE->db->count_all($table);
   $offset  = 0;

It should be around line 415. No comment out the code beginning with:

/*
if ($count > 0) //around line 421

... all the way through ...
$offset = $offset + $batch; // around line 465
}
*/

Then upload this to your test server and perform the upgrade per the instructions.

What this will do is convert your tables to UTF-8 collation but not the data in those tables. Which is OK, because using Polish and Russian, the data is already UTF-8 and we don’t want to double convert.

I look forward to your reply!

Cheers,

Thanks, Dan! I will try ASAP and report back.

Have a great weekend… Nico