forked from reactive-tech/kubegres
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Backlog
39 lines (24 loc) · 1.57 KB
/
Backlog
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
Features in order of priority:
- being able to configure how sensitive is hte automatic recovery process
by setting a time value in seconds
- #12 : Reuse PVC (see below) and Primary becomes a Replica
As part of the available options for the field "failover.pvc", there would be:
- "retain": the default option currently with Kubegres where PVC are kept but not reused for safety and investigation reasons
- "delete": the PVC will be deleted
- "reuse": if the state of the PVC is healthy, it will be reused by the newly created Replica pod. I think that matches with your suggestion?
- #20 : Delete PVC (is it the same as #12?)
- add a test in spec_readinessProbe_test and spec_livenessProbe_test so that
we test the default behaviour when no option is set from the YAML
- #30: Add one or many sidecar container options in Kubegres YAML
- #51: add documentation about how to recover backup
- add use cases documentation, for example how to expand storage manually and how to upgrade Postgres major version.
- check how to setup log archiving in case of replica does not found a data
- #46: Define Service Type for Primary and Replica
- #7 : Allow major version upgrade using pg_upgrade
- #35 : Restore database from a PV backup
- #82 : Allow using a unique image for the backup job
- #10 : Deploy Kubegres with a HELM chart
- #?: PG bouncer
- #? : Add a field to allow restarting StatefulSets and Pods via the YAML of "Kind: Kubegres"?
Blocked:
#49 : Expand Storage (waiting on the Kubernetes feature: https://github.com/kubernetes/enhancements/pull/2842)