From: Aleš Mrázek Date: Wed, 20 Sep 2023 13:21:41 +0000 (+0200) Subject: doc/upgrading.rst: better reference upgrading-to-6 X-Git-Tag: v6.0.5~19^2 X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=refs%2Fenvironments%2Fdocs-develop-doc-d160o2%2Fdeployments%2F2802;p=thirdparty%2Fknot-resolver.git doc/upgrading.rst: better reference upgrading-to-6 --- diff --git a/doc/upgrading-to-6.rst b/doc/upgrading-to-6.rst index 7f2ce8d10..f8247d031 100644 --- a/doc/upgrading-to-6.rst +++ b/doc/upgrading-to-6.rst @@ -24,7 +24,7 @@ So you can control the resolver using this systemd service. There is no need for managing multiple instances of ``kresd@.service`` like before version 6. However, ``kresd`` processes still run in the background as separate workers and are managed by new process ``knot-resolver-manager``. -Number of ``kresd`` workers can be configured directly in the new declarative configuration file. +The number of ``kresd`` workers can be configured directly in the new declarative configuration file. Knot Resolver's new configuration is by default located in ``/etc/knot-resolver/config.yaml``. .. code-block:: yaml @@ -55,14 +55,14 @@ and the equivalent option will very likely be in the :ref:`new declarative confi The documentation structure is basically the same. Otherwise, you will have to search for the option in the documentation separately. -If you have some custom Lua code in your configuration, you can use it in :ref:`lua section ` of declarativ configuration. +If you have some custom Lua code in your configuration, you can use it in :ref:`lua section ` of declarative configuration. However, it has some limitations and we cannot guarantee 100% functionality. For example, a configuration based on the systemd instance name will not work. Reconfiguration --------------- -To load the modified configuration without, just use ``reload`` and all running workers should be reconfigured without the resolver downtime. +To load the modified configuration, just use ``reload`` and all running workers should be reconfigured without the resolver downtime. This was not possible before version 6. It was necessary to manually restart all running ``kresd@`` instances. .. code-block:: bash diff --git a/doc/upgrading.rst b/doc/upgrading.rst index 4aa409c6e..a2d721780 100644 --- a/doc/upgrading.rst +++ b/doc/upgrading.rst @@ -29,10 +29,11 @@ newer versions when they are released. .. _`supervisord`: http://supervisord.org/ -5.x to 6.0 +5.x to 6.x ========== -* `detailed upgrade guide ` +* see the more detailed guide for `upgrading to version 6.x ` + 5.4 to 5.5 ==========