Gaining space on an install in mailing list mode

[Jose_C_Gomez]

Trying to recover some space on our server and our email_log table is 2GB
We use mailing list mode so we send approx 2 mil emails a month

Just wondering if there is a supported way to clean this up? I saw a few posts about clearing the email_log table but that can have some repercussion in usability?

Is there a more appropriate place to clean up? I have trimmed our backups down to 1 and that helped. But I need to recoup several gigs.
I am trying to get my storage upgraded but at the moment that isn’t a choice I can easily make.

Any other areas where I can clean up? I have done apt-get autoremove, and autopurge and launcher clean

Thanks!

[Falco]

Did you check the site setting delete email logs after days ?

[Jose_C_Gomez]

Yes that’s set to 30 but I still have 2 GB in this table. Which I guess I can live with just trying to find other places to trim.

[Falco]

I think that this will delete old email logs. I’d take a backup first, and wait for someone who knows what problems this might cause weighs in.

cd /var/discourse
./launcher enter app
rails c
EmailLog.where("DATE(created_at) < ?" , 1.month.ago).destroy_all

That is exactly what the setting does, no need to run a console command.

Yes that’s set to 30 but I still have 2 GB in this table. Which I guess I can live with just trying to find other places to trim.

Set it to 15? Maybe 10?

After that, you will need to expand storage in the server.

[Jose_C_Gomez]

So I currently have 26GB free which is great, however I’m still ruinning the old PostGre version. Have been having Perf issues lately so I really want to upgrade PostGre but it requires “double” the storage space to do that.
I’ve been limping along with the old one, but I was hoping to get upgraded. I need like a couple of gigs to meet the “double” threshold required by postgre for upgrade.

[Jose_C_Gomez]

Been running the 9.5 postgre template for like 6 months trying to find space… because of the nutty double space requirement. But lately my discourse slows to a crawl for no apparent reason, I don’t know what the issue is but I was hoping to eliminate that as an option by upgrading
sigh

[pfaffman (Jay Pfaffman)]

because of the nutty double space requirement.

It’s not nutty, it’s just math. The only way to do the upgrade is to make a copy of the database in the new format.

It is a bother, though.

You could do something like move to a new server or add another disk partition (a Volume if you’re on DigitalOcean) to temporarily gain space.

[Jose_C_Gomez]

I’m leaning towards something to do with email. We send out a LOT of email and it seems like when we have a very active topic that’s when sh*t hits the fan.
But I haven’t been able to figure out where the bottle neck is. Usually I end up doing an app rebuild , reboot and it settles out.
But within a few days it does it again. I know this is somewhat unrelated to the OP but if anyone has an idea… I’d love some thoughts.

[pfaffman (Jay Pfaffman)]

You might need more sidekiq jobs. My guess is that it’s getting slammed sending all of those emails.

When it’s slow, have a look at /sidekiq. Also you might just need more RAM. How big is your database and how much RAM do you have?

[Jose_C_Gomez]

Currently on Digital Ocean I have

DB Size: 17GB
RAM: 3GB

[pfaffman (Jay Pfaffman)]

I’d recommend that you bump up to 8GB. And re-run discourse-setup after you do to adjust the memory settings in your app.yml.

[tobiaseigen]

You could also just rely more on offsite backups. I keep 3 local backups and the rest on amazon.