Lari Hotari
Lari Hotari
@liangjianwen @liguangcheng There's a possible mitigation for the "continuous traffic data between two sites" requirement in #16651 . It doesn't remove it completely since the subscription replication can happen only...
For anyone dealing with subscription replication issues, one thing to check is whether replication snapshots are timing out. Increasing the timeout threshold (`replicatedSubscriptionsSnapshotTimeoutSeconds=30` -> `replicatedSubscriptionsSnapshotTimeoutSeconds=60`) could help solve issues in...
> Hey, was there any update on this ? I'm using pulsar 2.10.x , and facing the same issue. I read through the related issues on this repo, but even...
@spesalvi Yes, that's possible. Just don't specify `externalZookeeperServerList` since there's a different purpose for that setting. Instead, use `userProvidedZookeepers`. https://github.com/apache/pulsar-helm-chart/blob/afca5aaf08c02d011f7f8eaadfb19634eab6f6bb/charts/pulsar/values.yaml#L898-L901
Reopening this since there seems to be changed which cause userProvidedZookeepers not to work as expected.
It looks like the gap is about initializing the cluster metadata automatically. That only happens if you deploy the Zookeeper component. You will need to initialize the cluster metadata manually....
> **Pulsar Go Client version**: 0.12.0 In the [other comment](https://github.com/apache/pulsar/issues/24231#issuecomment-2865721960), you said that this problem is fixed with 0.15.0. Please don't report issues that have been fixed, it's preferred to...
@poorbarcode Do you have a chance to take a look at this since this is PIP-344 related?
> 3\. The callback methods can be executed in different thread everytime when we call `thenComposeAsync/thenApplyAsync/thenAcceptAsync`, it maybe lead to potential thread safety issue. As long as the mutations don't...