- 06 3月, 2017 2 次提交
- 04 3月, 2017 1 次提交
-
-
由 jai1 提交于
* Fix for Infinite loop in PersistentReplicator.startProducer() * Fix for Infinite loop in PersistentReplicator.startProducer()
-
- 03 3月, 2017 3 次提交
-
-
由 maurice barnum 提交于
Besides 500, other errors indicate that the server is having problems, so close the connection to avoid sending requests to the same bad server. DefaultKeepAliveStrategy will look at the request and response headers for other reasons to close, such as "Connection: close" headers.
-
由 Rajan 提交于
* Handle unknown exception on authorization check * fix test:init AuthorizationManager on Auth-enable and set status-500 on topic-lookup-internal-failire * fail partitionMetadata on zk-failure
-
由 Matteo Merli 提交于
If there is an error in reading from ZK, DiscoveryService should fail when reading partitions metadata (#272)
-
- 02 3月, 2017 10 次提交
-
-
由 Yuki Shiga 提交于
-
由 jai1 提交于
-
由 sschepens 提交于
-
由 Rajan 提交于
* close http-connection on internal server error from broker * broker kills current session on internal-server-error * add broker-address into response for client-debugging
-
由 Matteo Merli 提交于
-
由 Matteo Merli 提交于
-
由 jai1 提交于
-
由 Matteo Merli 提交于
-
由 jai1 提交于
-
由 Matteo Merli 提交于
-
- 01 3月, 2017 2 次提交
-
-
由 Matteo Merli 提交于
-
由 Rajan 提交于
-
- 28 2月, 2017 2 次提交
- 26 2月, 2017 2 次提交
-
-
由 Matteo Merli 提交于
-
由 Rajan 提交于
-
- 25 2月, 2017 7 次提交
-
-
由 Rajan 提交于
On batch-msg dispatching: broker should disconnect consumer which doesn't support batch-message (#215) * On batch-msg dispatching: broker should disconnect consumer which doesn't support batch-message * Close consumer at broker-side without closing connection if consumer doesn't support batch-message * Fail unsupportedBatchVersion-subscriber if topic has served batch-message
-
由 jai1 提交于
-
由 jai1 提交于
Corrected test case ProxyAuthentication, ProxyPublishConsumeTls and ProxyPublishConsumeWithoutZK (#245)
-
由 Rajan 提交于
Avoid triggering startReplProducer on newAddProducer as it may flips replicator state wrongly (#232) * Avoid triggering startReplProducer on newAddProducer as it may flips replicator state wrongly * Signal replicator is stopping if porducer is not created yet * read repl-cluster from policies to avoid restart of closing-replicator
-
由 Rajan 提交于
-
由 jai1 提交于
-
由 Yuki Shiga 提交于
* Make C++ tls auth api same as Java * Add tls auth test for Java client * Add tls auth test for C++ client
-
- 24 2月, 2017 3 次提交
-
-
由 jai1 提交于
-
由 Matteo Merli 提交于
-
由 Rajan 提交于
-
- 23 2月, 2017 1 次提交
-
-
由 jai1 提交于
Using TLSv1 since boost1_41 supports TLS v1.0
-
- 22 2月, 2017 1 次提交
-
-
由 Rajan 提交于
-
- 18 2月, 2017 6 次提交
-
-
由 jai1 提交于
-
由 Rajan 提交于
* Derive direct memory used count from netty counter * Fix: avoid initial direct-memory assertion with 0
-
由 jai1 提交于
-
由 nkurihar 提交于
* Fix #199: Do not stop to replicate when producer throws exception * Fix log messages * testReplicatorProducerClosing shoud be executed at last since it closes pulsar2/pulsar3 * Add unit test for replication resumption on backlog exceeded
-
由 Matteo Merli 提交于
-
由 Jai ASher 提交于
-