From 4cebc10f3cdc3eba904cd5d20441dc6223098bdf Mon Sep 17 00:00:00 2001 From: Daan Hoogland Date: Fri, 12 Dec 2025 15:40:11 +0100 Subject: [PATCH 1/2] guard explanation --- source/upgrading/index.rst | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/source/upgrading/index.rst b/source/upgrading/index.rst index 11631b56b0..43b9bf557b 100644 --- a/source/upgrading/index.rst +++ b/source/upgrading/index.rst @@ -34,6 +34,15 @@ The `cloudstack-management` package will now include the systemVM Templates for either prior upgrade or during fresh installation, ACS will handle the Template registration automatically, by mounting the secondary store onto the management server, copying the respective Templates to the store and then creating the `template.properties` file. +From ACS 4.20.3 and 4.22.1 onwards A starting MS that detects an +upgrade is needed but also sees mshost entries for MSses that are up, +will abort. If as an operator starting an MS you get the error +`Database upgrade is required but the management server is running in +a clustered environment.` during upgrade, check if you have stopped +all MSses in your environment and retry. If is still fails it might be +that an MS has stopped un-gracefully and you need to fix the DB-entry +for it. + .. note:: For information on the API changes and issues fixed in this release, please see the Release Notes section of the documentation From 9fa54faf69b0f2f3e0b58a82e3adcba7e93acd49 Mon Sep 17 00:00:00 2001 From: dahn Date: Wed, 17 Dec 2025 14:03:38 +0100 Subject: [PATCH 2/2] Add example of trouble shooting --- source/upgrading/index.rst | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/source/upgrading/index.rst b/source/upgrading/index.rst index 43b9bf557b..5372f3b325 100644 --- a/source/upgrading/index.rst +++ b/source/upgrading/index.rst @@ -41,7 +41,9 @@ will abort. If as an operator starting an MS you get the error a clustered environment.` during upgrade, check if you have stopped all MSses in your environment and retry. If is still fails it might be that an MS has stopped un-gracefully and you need to fix the DB-entry -for it. +for it. For example, make sure all MSses are stopped by checking the +hosts in the output of `select name from mshost where state <> ‘Down’` +and then, when all is safe, run `update mshost set state = ‘Down'` .. note:: For information on the API changes and issues fixed in this release, please see the Release Notes section of the documentation