Best upgrade method from 5.0LTS to 7.0LTS

[zbx_nm]

Hi all,

First time doing a Zabbix upgrade at all, so looking for some advice.

We have a CentOS 7 box running Zabbix 5.0 LTS / MySQL 5.5
Our end state is Ubuntu 22 with Zabbix 7.0 LTS / MySQL whatever in the repo (currently 8)

I am trying to work out the best way to get this done.
Originally I was going to install Zabbix 5.0 on a new Ubuntu 22 box and migrate the DB and then repo upgrade it, but there is no server version for Ubuntu 22.
I then considered doing an Inplace upgrade from CentOS 7 to Rocky 8 and then doing a Zabbix upgrade to version 6 - which mostly worked.... but i didn't do the DB upgrade right and rolled back

I am reading that I can just do a fresh install of Zabbix 7 on Ubuntu and then just import the 5.0 database? But i don't get how that will work with the schema updates required in version 6

It was at this point i decided to just post here for some advice.

Another thing i am tossing around doing is migrating just the database to a temporary server and then doing the rest of the upgrade.

Have servers and VMs to spare. Know enough about linux to be comfortable with doing these kinds of things - i just want the method that will give us stability long term. We do have a few custom templates and things, but it is a pretty vanilla install.

Thoughts appreciated, or even blogs. I did read a few, but they differed wildly

TIA

[Markku]

Install new server(s) for Zabbix and database, I'd recommend Debian 12 but you go with what you like. Then dump and import the old database to the new database server, and start Zabbix server with it, it will update the database automatically.

Remember to read and understand all the release and upgrade notes between 5.0 and 7.0. There may be some database changes that you should/must do manually as well.

This way you have no limits with the retries if needed.

Markku

[cyber]

Yep, as Markku says... just import your old DB schema and data to new DB, first start of zabbix server should do all upgrade tasks. Remember not to install new schema before importing DB.

[zbx_nm]

Thanks. I'll give that method a go then

[zbx_nm]

Worked a treat. Guess I was overthinking it just a bit haha

The only issue i had was with some locales, but those issues are documented in other posts.

For anyone finding it later, remember to tweak your InnoDB settings in MySQL / MariaDB BEFORE starting the import of a 44GB SQL file (facepalm)

[Terredise]

Worked a treat. Guess I was overthinking it just a bit haha

The only issue i had was with some locales, but those issues are documented in other posts.

For anyone finding it later, remember to tweak your InnoDB settings in MySQL / MariaDB BEFORE starting the import of a 44GB SQL file (facepalm)

hi, I've got a 30G SQL dump from MariaDB waiting to import to MySQL on a new VM. Not a DB guru, so would you please elaborate on tweaking InnoDB settings? And I assume you just created a new Zabbix DB and used the command mysql -uroot -p zabbix < zabbix.sql? Thank you.

[idanam7761]

yeah, second to the elaboration on tweaking InnoDB settings.

[zbx_nm]

Sorry, I only just saw these and probably well outdated now.
For me I didn't tweak anything before the import so it was hamstrung.

I ended up using some online calculators and ended up with this, however it may not be right your for your deployment and encourage you to understand the settings before you use this:

[mysqld]
host_cache_size=0
skip-name-resolve=1
max_connections=300
join_buffer_size=256MB
innodb_buffer_pool_size=4G
innodb_log_buffer_size=512M
innodb_log_file_size=1G
innodb_write_io_threads=16
innodb_buffer_pool_instances=8
innodb_redo_log_capacity=1G
innodb_flush_log_at_trx_commit=0
symbolic-links=0

[cyber]

innodb_file_per_table=ON
I have been once suggested to turn this on...
https://dev.mysql.com/doc/refman/9.1...blespaces.html