Karthik, to consume what was produced prior to subscription you need to use a reader, not a consumer. It is one of the perf admin tools that does not take a subscriber param. It won't drain any messages that have a subscriber backlog tied to them however.
On Tue, Apr 10, 2018 at 5:11 AM, Apache Pulsar Slack < apache.pulsar.sl...@gmail.com> wrote: > 2018-04-09 17:10:50 UTC - Karthik Palanivelu: Hi All, I tested the above > scenario. Only global cluster gets replicated. How could I achieve the same > in topics created local to cluster in different datacenter's? Because, when > us-east-1 of aws is unavailable, I should be having my application running > from us-west with the same set of data with no data loss. Please advice > ---- > 2018-04-09 17:13:13 UTC - Karthik Palanivelu: Hello all, Can you please > let me know how I can consume all the messages that are not consumed yet? > For example, I produced x set of messages with no consumers. After > producing messages, when client comes online as Shared/Exclusive not > consuming the messages. It consumes messages that were produced after > consumer is live. Please let me know if I am missing any. > ---- > 2018-04-09 17:27:30 UTC - Karthik Palanivelu: @Karthik Palanivelu uploaded > a file: <https://apache-pulsar.slack.com/files/U7VRE0Q1G/FA3DN78KE/ > -.txt|Untitled> > ---- > 2018-04-09 17:28:10 UTC - Karthik Palanivelu: This seems like I need to > have all my topics in global namespace. > ---- > 2018-04-10 00:48:13 UTC - Sijie Guo: @Karthikeyan Palanivelu > > yeah only global namespaces are able to configure cross-datacenter > replications. if you want to have a namespace replicated to multiple > datacenters, you need to create the namespace under global. “global” > cluster just means the replication configuration is stored globally, so > every data center is in-sync with replication settings. but you can still > configure namespaces to only replicated between specific clusters. > > “global” vs “local” are kind of legacy concept when pulsar was intiailly > developed at yahoo. at the coming pulsar 2.0 release, we are removing > “cluster” concept. > > but for now, using 1.2x version, you need to make the namespace under > “global” scope in order to replication. > ---- > 2018-04-10 00:49:06 UTC - Sijie Guo: > Hello all, Can you please let me > know how I can consume all the messages that are not consumed yet? For > example, I produced x set of messages with no consumers. After producing > messages, when client comes online as Shared/Exclusive not consuming the > messages. It consumes messages that were produced after consumer is live. > Please let me know if I am missing any. > ---- > 2018-04-10 00:50:53 UTC - Sijie Guo: > Hello all, Can you please let me > know how I can consume all the messages that are not consumed yet? For > example, I produced x set of messages with no consumers. After producing > messages, when client comes online as Shared/Exclusive not consuming the > messages. It consumes messages that were produced after consumer is live. > Please let me know if I am missing any. > > if a subscription is first time, the consumer starts from the latest > message (by default) at the point that subscription was created. You can > configure to start from earliest message. Or you can rewind the > subscription back at time using bin/pulsar-admin. > ---- > 2018-04-10 06:55:54 UTC - Mate Varga: @Mate Varga has joined the channel > ---- >
