Pod is very slow in producing a wall. How to fix?

Dear community,

The pod i’m with has become very slow over the years, taking up to 15 seconds to load a “wall”.
(but personal posts-page loads fast)

i spoke to the pod-admin;
They use a gigibit internet connection, and SSD disks. runs D* version [0.7.14.0]
He did some SQL housekeeping, but with no performance improvements.
The ibdata -table has grown over the years to be 40GB, and he suspects that to be the culprit.

Any idea’s, Docs etc. on maintenance getting the server fast again are welcome.
Thanks

i had an answer via IRC

[5:32:54 PM] hmn: ibdata sounds like mysql, diaspora doesn’t scale well on mysql, that’s why postgres is recommended. So you probably should look into migrating to postgres.
[5:34:44 PM] but also, 40GB sounds like really a lot, do you have that many users with that many contacts? Or did you enable the relay at one point which floods your pod with a lot of posts?

Question stays…

How to make the ibdate table smaller again?

It would probably help if the podmin themselves communicates with support, either in this thread or via IRC. They will be able to give information about the pod setup and the database that will help to troubleshoot it.

1 Like

@r-hmn It seems ibdata grows over time since it stores write logs and temp thing that don’t get automatically cleaned up. Percona is a commercial version of MySQL and they have an article with more detail here: https://www.percona.com/blog/2013/08/20/why-is-the-ibdata1-file-continuously-growing-in-mysql/

Here’s a doc on cleanup after an upgrade, though it will probably work in the general case. They’re basically dumping and rebuilding the database. API Gateway: Cleaning up a large ibdata file after upgrade to 9.3 and newer.

If you spend a bit more time searching, you may find something more directly helpful.

Thanks! i will forward this to the podmin