MyISAM to InnoDB conversion takes forever

[lucakuehne]

Hi

I'm planning on updating to version 6.0.0 of EspoCRM.
For that I wanted to change the tables from MyISAM engine to InnoDB.
But the script has been running since 21 hours, and most of the time for the "emails" table.

I don't know if the script is still running.
EspoCRM is since then not really usable, because all the emails are not showing up. So our sales team is not able to work.

How can we solve that, because if it is going to run for 50 or more hours nobody is able to work for the rest for the week?
Or how can we perform this change faster?

Kind regards
Luca

[Maximus]

Hi there,
It shouldn't take so much time to convert. It looks lite the process has failed. Do you use this script to change a DB engine https://www.espocrm.com/blog/convert...ine-to-innodb/ ?

Please check an EspoCRM log and your DB error log files.

[lucakuehne]

Hi

I checked the logs of EspoCRM and there was absolutely no entry after the update started...
I then aborted the command, and restarted mysql as well as apache.

Then I did start de conversion of the email table to InnoDB from the mysql cli, and it has been running since 13.5h and it shows as the progress 49.6%.
So this changes really seems to take forever.

I guess there is no way to speed that up, or is there a way?

[Maximus]

You didn't answer - how exactly you are doing the conversion (manually by running a query like ALTER TABLE `emails` engine=InnoDB, or by executing our script)?
How many emails records do you have in the 'email' table?

[lucakuehne]

Sorry...
For the first try I tried doing it using your script: https://www.espocrm.com/blog/convert...ine-to-innodb/

For the second try I started yesterday in the evening, I'm using the mysql cli using the following command, because there I can see the progress:

ALTER TABLE `emails` engine=InnoDB

It has now been running for 19h and the progress is at 65.6%.

Our email table has 114970 records and is 4,8 GiB in size.

[Maximus]

Hi,
Sorry for the delayed reply.
It is weird. 114970 records it is not a big number of records. It shouldn't take so much time. It looks like you have some issue with your DB. Or your DB stands on the very slow machine.