Changes since last release: v3.0.11...v3.0.12
- New analysis:
UnreachableMasterWithLaggingReplicas
, handles situations where the master is in lockdown and replicas still think everything is fine. Handles by restarting IO thread on replicas, #572 - Massive reduction and redaction of logs, #555
- GTID-based promotion utilize immediate master promotion optimization, previously designed for Pseudo-GTID based failovers, #551
- Docker builds now using current branch as opposed to pulling master, #550
- Docker image now smaller, thanks @rhoml, #545
- Fix to basic authentication in
orchestrator-client
, #581 - Graceful master takeover: fail when attempting to promote ignored host, #570
- Graceful master takeover: now works when master is
read-only
, #548 - Added
FailMasterPromotionIfSQLThreadNotUpToDate
configuration, thanks @samiahlroos, #534 orchestrator-client
uses incremental sleep to auto-recover brief leader loss, #527- Fixed timestamp on
sqlite
in non UTC boxes, #591, thanks @ndelnano start-replica
returns an error if replication does not kick in in timely manner, #590, thanks @cezmunsta- Fixed
take-master
logic when server already is co-master, thanks @Pomyk, #576
There's much more, and there are outstanding PRs which I am yet to review/test. Thank you to the many contributors!
$ gsha256sum *
2fbe927d9ee725d80ff30ba85dba2139c33bde5007f388ede8f4a05ee00799c1 orchestrator-3.0.12-1.x86_64.rpm
918d6950616b2cf16db019c7af8c7dae223de7b7ff1e3e221cad34baf734b0bb orchestrator-3.0.12-linux-amd64.tar.gz
de8e639636d9d7fe17062780d4ad67ae43a9e0d2b913c62bac347aeffcc57dd4 orchestrator-cli-3.0.12-1.x86_64.rpm
9252b0cd387ea090c01983cb59cb510ad9d725236f3ef6adabff9d3a81ae456a orchestrator-cli_3.0.12_amd64.deb
8aeb69c35225e30692ea3c203955314b0571e36539af64c92c76b0c59c745b0a orchestrator-client-3.0.12-1.x86_64.rpm
ec8558217103be7a2457aabc6663545757332192f1e71310b59bed3a784fea53 orchestrator-client_3.0.12_amd64.deb
653ae7b6e2edfffd50cced387ce6d60247ff287fa7a8866d54bceb88d1bcdca6 orchestrator_3.0.12_amd64.deb