XF 2.1 mail_queue (4gb) and reaction_content (2gb) are my biggest two tables

HJW

Active member
On a forum with around 1m posts these are my biggest two tables.

Seems strange that mail queue is twice the size of post and reaction content is the same size.

3993616 /var/lib/mysql/xf/xf_mail_queue.ibd
2117648 /var/lib/mysql/xf/xf_reaction_content.ibd
2011160 /var/lib/mysql/xf/xf_post.ibd
1998860 /var/lib/mysql/xf/xf_ip.ibd

Now I'm running out of space I need to clear some of them, is there something I should be running to clean up some of these tables?

My whole DB gzipped is less than 2gb but a search on the server says its 14gb

I've tried optimising these tables but it seems to make no change, is there something else I should be doing?
 
Last edited:
yep mail queue empty for me too but still 4gb :LOL:

reaction content is 11,970,041 rows and 2 Gb

Not sure if that's right, I'd assume not as 1m posts is about the same size
 
yep mail queue empty for me too but still 4gb :LOL:

reaction content is 11,970,041 rows and 2 Gb

Not sure if that's right, I'd assume not as 1m posts is about the same size

You probably need to optimize that table using the OPTIMIZE TABLE syntax or in PHPMyAdmin > Operations > Table Maintenance > Optimize table
 
You probably need to optimize that table using the OPTIMIZE TABLE syntax or in PHPMyAdmin > Operations > Table Maintenance > Optimize table
Tried an optimize before and didn't do anything.

Tried a defrag just now and that fixed it :D

Would it have been the combination of a defrag and optimise that fixed it?

I want to try it on reaction content and IP but they both have 10+ million rows, so would it be a better idea to turn off the forum then do it?
 
well that went well trying to optimise the search index table :eek: and the site offline for a bit

1591190537823.webp

I guess because it's MyISAM it can't be defragged? No difference to the size and 1.3 gb with 1.4 rows, which probably is about right

Now scared to do the reaction_content table incase the site is down for an hour

so defrag is essentially taking a mysqldump of the database and importing it into a fresh database?
I wouldn't of thought so, but I know nothing :LOL:
 
might be worth considering. stop the forum. extract forum database to sql file. import sql file to a new database. see if the new database has same issues? if not, switch to that database?
 
  • Like
Reactions: HJW
your server seems pretty powerful. pausing the forum is a must. extracting the database and then importing it into a fresh one should not take more than 30 minutes :)
 
The mail xf_mail_queue one has been fixed with a defrag. Seems to need to be done pretty regularly as it fills up to hundreds of mb even though its empty

But reaction content is now 5gb and 15 million rows and I've dare not run a defrag on that yet :oops: it's been on my list to do..
 
@HJW MySQL table compression works well at dropping disk-usage, it does increase memory pressure a little. But if you can OpenZFS's inline compression is amazing for MySQL databases. The ability to make cheap snapshots is also great for database backups.

The mail xf_mail_queue one has been fixed with a defrag. Seems to need to be done pretty regularly as it fills up to hundreds of mb even though its empty
This is somewhat expected, as all content is journaled to the queue before being sent. This makes sending notifications a lot speedier, but has this trade-off.
 
Last edited:
  • Like
Reactions: HJW
By defrag do you mean by running the optimize command, or something else?
This:

1611675237771.webp

@HJW MySQL table compression works well at dropping disk-usage, it does increase memory pressure a little. But if you can OpenZFS's inline compression is amazing for MySQL databases. The ability to make cheap snapshots is also great for database backups.
Thanks that's interesting I'll look into it. Currently getting 5million reactions a month, so the reactions table is growing quickly!
 
Back
Top Bottom