hydra: 2020-02-06 -> 2020-03-{24,27}

Upgrades Hydra to the latest master/flake branch. To perform this
upgrade, it's needed to do a non-trivial db-migration which provides a
massive performance-improvement[1].

The basic ideas behind multi-step upgrades of services between NixOS versions
have been gathered already[2]. For further context it's recommended to
read this first.

Basically, the following steps are needed:

* Upgrade to a non-breaking version of Hydra with the db-changes
  (columns are still nullable here). If `system.stateVersion` is set to
  something older than 20.03, the package will be selected
  automatically, otherwise `pkgs.hydra-migration` needs to be used.

* Run `hydra-backfill-ids` on the server.

* Deploy either `pkgs.hydra-unstable` (for Hydra master) or
  `pkgs.hydra-flakes` (for flakes-support) to activate the optimization.

The steps are also documented in the release-notes and in the module
using `warnings`.

`pkgs.hydra` has been removed as latest Hydra doesn't compile with
`pkgs.nixStable` and to ensure a graceful migration using the newly
introduced packages.

To verify the approach, a simple vm-test has been added which verifies
the migration steps.

[1] https://github.com/NixOS/hydra/pull/711
[2] https://github.com/NixOS/nixpkgs/pull/82353#issuecomment-598269471
This commit is contained in:
Maximilian Bosch
2020-03-28 23:33:25 +01:00
parent ea5c800175
commit bd5324c4fc
9 changed files with 444 additions and 231 deletions
@@ -697,6 +697,66 @@ auth required pam_succeed_if.so uid >= 1000 quiet
</para>
</warning>
</listitem>
<listitem>
<para>
<package>Hydra</package> has gained a massive performance improvement due to
<link xlink:href="https://github.com/NixOS/hydra/pull/710">some database schema
changes</link> by adding several IDs and better indexing. However, it's necessary
to upgrade Hydra in multiple steps:
<itemizedlist>
<listitem>
<para>
At first, an older version of Hydra needs to be deployed which adds those
(nullable) columns. When having set <link linkend="opt-system.stateVersion">stateVersion
</link> to a value older than <literal>20.03</literal>, this package will be selected
by default from the module when upgrading. Otherwise, the package can be deployed using
the following config:
<programlisting>{ pkgs, ... }: {
<link linkend="opt-services.hydra.package">services.hydra.package</link> = pkgs.hydra-migration;
}</programlisting>
</para>
</listitem>
<listitem>
<para>
Automatically fill the newly added ID columns on the server by running the following
command:
<screen>
<prompt>$ </prompt>hydra-backfill-ids
</screen>
<warning>
<para>Please note that this process can take a while depending on your database-size!</para>
</warning>
</para>
</listitem>
<listitem>
<para>
Deploy a newer version of Hydra to activate the DB optimizations. You can choose from
either <package>hydra-unstable</package> (latest <literal>master</literal> compiled
against <package>nixUnstable</package>) and <package>hydra-flakes</package> (latest
version with flake-support).
<warning>
<para>
If your <link linkend="opt-system.stateVersion">stateVersion</link> is set to
<literal>20.03</literal> or greater, <package>hydra-unstable</package> will be used
automatically! This will break your setup if you didn't run the migration.
</para>
</warning>
Please note that Hydra is currently not available with <package>nixStable</package>
as this doesn't compile anymore.
</para>
</listitem>
</itemizedlist>
<warning>
<para>
<package>pkgs.hydra</package> has been removed to ensure a graceful database-migration
using the dedicated package-attributes. If you still have <package>pkgs.hydra</package>
defined in e.g. an overlay, an assertion error will be thrown. To circumvent this,
you need to set <xref linkend="opt-services.hydra.package" /> to <package>pkgs.hydra</package>
explicitly and make sure you know what you're doing!
</para>
</warning>
</para>
</listitem>
</itemizedlist>
</section>