plugin-barman-cloud/internal
Marco Nenciarini bb0d217941
fix: avoid rolling out recovery-only pods just to drop the sidecar
Gating sidecar injection purely on cluster initialization state made
the operator's periodic EVALUATE drift-check disagree with the pod
spec stored at creation time, so it would roll out the primary right
after it finished bootstrapping, solely to strip the now-unneeded
sidecar.

Only stop adding the sidecar on TYPE_CREATE, once the cluster is
initialized; keep injecting it unconditionally on TYPE_EVALUATE so an
already-running pod never drifts against its own stored spec.

Signed-off-by: Marco Nenciarini <marco.nenciarini@enterprisedb.com>
2026-07-27 17:47:05 +02:00
..
cmd fix(restore): use custom CNPG group and version for scheme registration (#847) 2026-04-13 13:20:38 +02:00
cnpgi fix: avoid rolling out recovery-only pods just to drop the sidecar 2026-07-27 17:47:05 +02:00
controller fix(deps): update module sigs.k8s.io/controller-runtime to v0.24.0 (#880) 2026-05-06 10:15:29 +02:00
scheme fix(scheme): register meta types for CNPG scheme group version (#943) 2026-06-08 11:08:15 +02:00