* Handle retry connect to Postgres when ping return EOF error.
* Update pkg/cluster/database.go
Co-authored-by: Felix Kunde <felix-kunde@gmx.de>
---------
Co-authored-by: Trung Minh Lai <trung.lai@hitachivantara.com>
Co-authored-by: Felix Kunde <felix-kunde@gmx.de>
* add prefix /vol- on when EBS doesn't have
* add new unit test for to get the volumeID
* add a prefix to search in the string of volumeID
---------
Co-authored-by: Jociele Padilha <jociele.padilha@zalando.de>
* bump to v1.9.1
* update year in license and add links to more blog posts
* bump go to 1.19 and update dependencies
* go for 1.10.0 instead of 1.9.1
* fix unit test - removed obsolete ClusterName field
* fix DNS template in UI helm chart deployment file
* bump pooler image
* set pooler pod security context
* use hard coded RunAsUser 100 and RunAsGroup 101 for pooler pod
* unify generation of TLS secret mounts
* extend documentation on tls support
* add unit test for testing TLS support for pooler
* add e2e test for tls support
* fix to pooler TLS support, security context fsGroup added (#2216)
* add environment variable of CA cert path in pooler pod template
* additional logic for custom CA secrets and mount path
* fix ca file name
* Introduce `masterServiceAnnotations` & `replicaServiceAnnotations`
Introduce `masterServiceAnnotations` & `replicaServiceAnnotations` to the `Postgresql` CRD.
`masterServiceAnnotations` overrides `serviceAnnotations` for master role if not empty.
`replicaServiceAnnotations` overrides `serviceAnnotations` for replica role if not empty.
Existing definition of `serviceAnnotations` continue to work for backward compatibitlity when neither `masterServiceAnnotations` nor `replicaServiceAnnotations` is defined.
This closes https://github.com/zalando/postgres-operator/issues/1927
* Accumulate service annotations
First, global config, then ServiceAnnotations overriding, then MasterServiceAnnotations and ReplicaServiceAnnotations.
This addresses
https://github.com/zalando/postgres-operator/pull/2161#discussion_r1063558711.
* Update admin doc with master & replica service annotations overrides
Addressed https://github.com/zalando/postgres-operator/pull/2161#discussion_r1064744086
Co-authored-by: Felix Kunde <felix-kunde@gmx.de>
* add volume with custom TLS config to pooler deployment
* bump pg bouncer image tag which support new feature
Co-authored-by: Jérémie Seguin <jeremie.seguin@malt.com>
* new config options to specify resources for logical backup jobs
* bug in logical backup script for s3 dumps
* define enum for logical_backup_provider
* changed order of logical backup azure options
* fix unit test for stream comparison
* Bumped Spilo image tag to the one that supports PostgreSQL 15. Using CDP version temporarily until non-CDP one is released.
* Added support for PostgreSQL 15 and made it default. 9.5 and 9.6 are now no longer supported
* Bumped spilo image tag to 2.1-p9
* Bumped spilo image in test launcher
Co-authored-by: yoshihiko <ariyoshi10@gmail.com>
Co-authored-by: Felix Kunde <felix-kunde@gmx.de>
* Allow drop slots when it gets deleted from the manifest
* use leader instead replica to query slots
* fix and extend unit tests for config update checks
Co-authored-by: Felix Kunde <felix-kunde@gmx.de>
If you import the zalando postgresql v1 api with kubebuilder, it
complains about the missing tags.
```
❯ make manifests
test -s /.../bin/controller-gen && /.../bin/controller-gen --version | grep -q v0.10.0 || \
GOBIN=/.../bin go install sigs.k8s.io/controller-tools/cmd/controller-gen@v0.10.0
/.../bin/controller-gen rbac:roleName=manager-role crd:allowDangerousTypes=true webhook paths="./..." output:crd:artifacts:config=config/crd/bases
/.../go/pkg/mod/github.com/zalando/postgres-operator@v1.8.2/pkg/apis/acid.zalan.do/v1/postgresql_type.go:116:2: encountered struct field "Everyday" without JSON tag in type "MaintenanceWindow"
/.../go/pkg/mod/github.com/zalando/postgres-operator@v1.8.2/pkg/apis/acid.zalan.do/v1/postgresql_type.go:117:2: encountered struct field "Weekday" without JSON tag in type "MaintenanceWindow"
/.../go/pkg/mod/github.com/zalando/postgres-operator@v1.8.2/pkg/apis/acid.zalan.do/v1/postgresql_type.go:118:2: encountered struct field "StartTime" without JSON tag in type "MaintenanceWindow"
/.../go/pkg/mod/github.com/zalando/postgres-operator@v1.8.2/pkg/apis/acid.zalan.do/v1/postgresql_type.go:119:2: encountered struct field "EndTime" without JSON tag in type "MaintenanceWindow"
Error: not all generators ran successfully
run `controller-gen rbac:roleName=manager-role crd:allowDangerousTypes=true webhook paths=./... output:crd:artifacts:config=config/crd/bases -w` to see all available markers, or `controller-gen rbac:roleName=manager-role crd:allowDangerousTypes=true webhook paths=./... output:crd:artifacts:config=config/crd/bases -h` for usage
make: *** [manifests] Error 1
```
This commit adds support of a not-yet-released Patroni feature that allows postgres to run as primary in case of a failed leader lock update.
* Add Patroni 'failsafe_mode' local parameter (enable for a single PG cluster)
* Allow configuring Patroni 'failsafe_mode' parameter globally
* add toggle to turn off readiness probes
* include PodManagementPolicy and ReadinessProbe in stateful set comparison
* add URI scheme to generated readiness probe
* create streams only after postgres instances were restarted
* checkAndSetGlobalPostgreSQLConfiguration returns if config has been patched
* restart can be pending even without a config patch
* deprecate ClusterName field of Postgresql type
* remove for teamId from operator API endpints /status /logs /history
* update dns_format_string and yaml template in UI
* Use getSwitchoverCandidate instead of masterCandidate when trying to migrating master pod to a replica
Ref: #1983
* Remove unused masterCandidate (replaced by getSwitchoverCandidate)
Ref: #1983
* allow in place pw rotation of system users
* block postgres user from rotation
* mark pooler pods for replacement
* adding podsGetter where pooler is synced in unit tests
* move rotation code in extra function
* removing inner goroutine in cluster.Switchover and resolve race between processPodEvent and unregisterPodSubscriber
* unlock mutex after handling event, now with non-blocking default case
* reverse membership for additional owner roles
* remove type RoleOriginSpilo
* use e2e images with cron_admin inside
* let operator resolve reversed membership
* make additional owner roles part of the sync user strategy
* add more context in the docs about additional_owner_roles
* return err if teams API fails with StatusCode other than 404
* add unit test for 404 at team members
Co-authored-by: Jociele Padilha <jociele.padilha@zalando.de>
Co-authored-by: Felix Kunde <felix-kunde@gmx.de>
* add the possibility to create a standby cluster that streams from a remote primary
* extending unit tests
* add more docs and e2e test
Co-authored-by: machine424 <ayoubmrini424@gmail.com>
* return only warning if team can't be found
Co-authored-by: Jociele Padilha <jociele.padilha@zalando.de>
Co-authored-by: Felix Kunde <felix-kunde@gmx.de>
* do not create endpoints when use config maps
* delete cluster objects with 'leader' suffix
Co-authored-by: Евграфов Александр Александрович <aevgrafov@cmx.ru>
Ida Novindasari
Felix Kunde
Thinking Chen
Trung Minh Lai
Jan Jansen
Jociele Padilha
drivebyer
genofire
Pavel Ven Gulbin
Dmitry Volodin
Owen Ou
jeremie-seguin
Stef Graces
yoshihikoueno
Francois Parquet
Christian Rohmann
Camilo Schöningh
Polina Bungina
machine424
Philipp B
Jan Mussler
JBWatenbergScality
mujx
Oleg
neelasha-09
evsasha