view mod_migrate/README.markdown @ 5119:048e339706ba

mod_rest: Remove manual reference expansion in schema This hack was originally added to reduce the number of definitions of common attributes (type, to, from etc) and payloads (e.g. delay). This predated pointers and references, and until now was needed because parsing picked out the correct stanza kind from the schema, which broke internal references. Removing this hack paves the way for allowing the schema to be configured or customized more easily.
author Kim Alvefur <zash@zash.se>
date Tue, 20 Dec 2022 21:48:28 +0100
parents c5122b7633a8
children
line wrap: on
line source

---
summary: prosodyctl cross storage driver migration tool
...

Introduction
============

This module adds a command to `prosodyctl` for copying data between
storage drivers.

Usage
=====

    prosodyctl mod_migrate example.com <source-store>[-<store-type>] <target-driver> [users]*

`<source-store>` would be e.g. `accounts` or `private`. To migrate
archives, the optional suffix `<store-type>` would be set to `archive`,
so e.g. `archive2-archive` or `muc_log-archive`. Multiple stores can be
given if separated by commas.

`<target-driver>` is the storage driver to copy data to, sans the
`mod_storage_` prefix.

`mod_migrate` tries to request a list of users from `usermanager`, but
this does not always work. If so, you can supply usernames as arguments
after the target driver.

The process is something like this:

1.  Decide on the future configuration and add for example SQL
    connection details to your prosody config, but don't change the
    `store` option yet.
2.  With Prosody shut down, run
    `prosodyctl mod_migrate example.com accounts sql`
3.  Repeat for each store, substituting 'accounts'. E.g. vcards,
    private...
4.  Change the [`storage` configuration](https://prosody.im/doc/storage)
    to use the new driver.
5.  Start prosody again.

Examples
========

``` sh
prosodyctl mod_migrate example.com accounts,roster,private,vcard sql
```

Compatibility
=============

Should work with 0.8 and later.