[etcd-backup-restore]
Action Required
- [USER] Defragmentation schedule can be configured now in cron standards using flag
defragmentation-schedule
. ⚠️ Removed the flagdefragmentation-period-hours
. (#194, @swapnilgm) - [USER] ⚠️ Removed the flag
delta-snapshot-period-seconds
. Instead use replacement flagdelta-snapshot-period
with input value format supported by golangtime.Duration
. (#194, @swapnilgm) - [USER] ⚠️ Removed the flag
garbage-collection-period-seconds
. Instead use replacement flaggarbage-collection-period
with input value format supported by golangtime.Duration
. (#194, @swapnilgm)
Most notable changes
- [DEVELOPER] Added TLS support for backup-restore server. TLS can be enabled by passing the paths to both TLS cert and key PEM-format files via
--server-cert
and--server-key
flags. (#196, @shreyas-s-rao)
Improvements
- [USER] Fix the error handling in revision consistency check leading to restoration on valid etcd data directory. (#200, @swapnilgm)
- [USER] Expose http API to trigger out-of-schedule delta snapshot. (#192, @swapnilgm)
- [OPERATOR] Fix the object listing for OSS snapstore (#197, @swapnilgm)
- [OPERATOR] [Fix] Cleanup in-memory events before stopping snapshotter. (#190, @swapnilgm)
- [OPERATOR] Added documentation for restoration. (#188, @shreyas-s-rao)
- [OPERATOR] Added negative tests for restoration of snapshots. (#173, @ashwani2k)
- [OPERATOR] Performance regression tests can be executed using the
make perf-regression-test
target against the Kubernetes cluster pointed to by the environment variable PERF_TEST_KUBECONFIG. (#171, @amshuman-kr) - [DEVELOPER] Expose new metric
etcdbr_snapshot_required
. (#193, @shreyas-s-rao)