Mattermost, Inc.

Mattermost fails connecting to pgpool?

While trying to run mattermost,
It’s working successfully with a single postgresql instance.

and while tries to run the same with postgresql cluster setup(for scaling)
It is failing to connect to database at some specific points and
throwing the below error.

[EROR] Failed to retrieve user profiles
team_id=5esim8ypjpbdub75hy693nwfny, err=SqlUserStore.GetProfiles:
store.sql_user.get_profiles.app_error, read tcp
x.x.x.x:49766->x.x.x.x:5432: i/o timeout

code
ref:https://github.com/mattermost/platform/blob/bb69e98631b2541954b6ae465b8ba5f788b9dc49/store/sql_user_store.go#L469

But at the same time chat, usercreation, and all are working properly on
the same setup. This is causing only while accessing the system
console.(attached a screen-shot of system console page)
,

and am not getting any clue to move forward, so please share if anyone
gets an idea and also please ask if more verbosity is needed

Thanks

Steps to reproduce the issue

Deploying on openshift
or
Deploying on docker
after deploying, do a signup and go to system console

Hi @syamgk,

Mattermost Team Edition is intended for self-hosted team communication, it’s not intended to be used at a scale where you’d need a database cluster. To officially support such a configuration would be too much of a tax on the open source community to setup such an environment and to test, document and support it properly, even if a change was submitted.

There’s a commercial extension, Mattermost Enterprise Edition, intended for large organizations that supports DB clusters (along with advanced access controls, compliance reporting, corporate directory integration and other features needed for advanced organizations beyond what teams need).

If you’re at an enterprise needing secure, modern messaging, it’s an excellent value. If you’re a non-profit or academic institution there’s steep discounts available.

You’re also welcome to fork the open source version and modify it as you wish–and many organizations do.

DB clustering is just out-of-scope for the Team Edition.