After upgrading to Pulsar 2.11.0, we encountered an issue where we
were unable to immediately query non-persistent topics that were
created in Pulsar. This behavior was unexpected as everything was
functioning properly prior to this release. We have identified two
known issues ([1] and [2]) that ar
I just treat the partitioned topic as a set of named non-partitioned
topics. Pulsar doesn't really have any partitioned definition. So the
non-partitioned topics would be the partition of the partitioned
topic.
I think the real confusion is that we can create the non-partitioned
topic independentl
.name("flink_table_schema")
> >> > //.type(SchemaType.BYTES)
> >> > //.schema(mapper.writeValueAsBytes(properties))
> >> > // SchemaType.JSON
> >> > .type(SchemaT
This feature is quite cool. I couldn't wait for it. It's quite useful
in the Flink connector. We can make the transaction with better
deduplication.
On Thu, Apr 6, 2023 at 8:30 PM Xiangying Meng wrote:
>
> Dear Pulsar Community,
>
> I am excited to invite you all to participate in the discussion
Hi all,
I want to emphasize that this issue
https://github.com/apache/pulsar/issues/19316 has not been fixed
either in 2.11.1. The newly introduced bug of non-persistent topic
query. This affects a lot of existing customers and should be put in
the highest priority.
Thanks
Yufan Sheng
On Mon
y use the schema when the users want
to do some evolution check. Otherwise, we will only send messages in
BYTES schema.
On Tue, Mar 28, 2023 at 10:06 AM SiNan Liu wrote:
>
> Hi yufan.
> Can you describe a bit the usage scenario of byte schema in
> flink-connector-pulsa?
>
>
>
As the flink-connector-pulsar developer, I don't want to disable the
BYTES schema upload. In my opinion, using BYTES schema means the users
want to bypass the schema check and handle the schema validation by
themselves.
On Tue, Mar 28, 2023 at 8:58 AM SiNan Liu wrote:
>
> Hi, everyone.
> When a u
