There are two problems I have encountered that cause kojid.service to fail.
The first is that if the service does not wait for the network connection to be online it cannot communicate with the server and fails to start.
The second is that gssapi requires accurate time. On systems with a good clock time being out of sync probably is not a big problem, however we haven't been so lucky thus far with riscv64 builders for https://riscv-koji.fedoraproject.org/koji/
Enabling chrony-wait.service and adjusting the kojid.service to have After=network-online.target httpd.service chrony-wait.service appears to result in reliable starts.
After=network-online.target httpd.service chrony-wait.service
I can't browse or create a PR from my fork otherwise I would submit a PR. Anyway, this is what is working for me:
From 8969c2650cee8b8813d49432ca03e44862fa7c32 Mon Sep 17 00:00:00 2001 From: Jason Montleon <jmontleo@redhat.com> Date: Fri, 8 Aug 2025 11:00:11 -0400 Subject: [PATCH] Improve reliability of kojid service start Signed-off-by: Jason Montleon <jmontleo@redhat.com> --- builder/kojid.service | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/builder/kojid.service b/builder/kojid.service index 0c63110c..3aaea44e 100644 --- a/builder/kojid.service +++ b/builder/kojid.service @@ -3,7 +3,7 @@ Description=Koji build server Documentation=https://docs.pagure.org/koji/server_howto/ # If the koji-hub is on this host, httpd.service is needed first -After=network.target httpd.service +After=network-online.target httpd.service chrony-wait.service [Service] ExecStart=/usr/sbin/kojid \ -- 2.50.1
Edit: I was able to create a PR: https://pagure.io/koji/pull-request/4442
Metadata Update from @mikem: - Custom field Size adjusted to None - Issue set to the milestone: 1.36
Metadata Update from @mikem: - Issue tagged with: no_qe
Commit b1e2e4b3 fixes this issue
This issue has been migrated to Fedora Forge: https://forge.fedoraproject.org/koji/koji/issues/4441
Please continue any further discussion there.