Update:
I can confirm my previous guess based on the changes in
https://issues.apache.org/jira/browse/FLINK-4280 that was merged for Flink
1.3.0.
When upgrading from Flink 1.2.x -> 1.3.0, the new startup position
configurations were respected over the checkpointed offsets (only once for
the first r
Hi Somya,
I'll have to take a closer look at the JIRA history to refresh my memory on
potential past changes that caused this.
My first suspection is this:
It is expected that the Kafka consumer will *ignore* the configured startup
position if the job was restored from a savepoint.
It will always
@gordon Do you remember whether we changed any behavior of the Kafka
0.10 consumer after 1.3.3?
On 23/01/2020 12:02, Somya Maithani wrote:
Hey,
Any ideas about this? We are blocked on the upgrade because we want
async timer checkpointing.
Regards,
Somya Maithani
Software Developer II
Helps
Hey,
Any ideas about this? We are blocked on the upgrade because we want async
timer checkpointing.
Regards,
Somya Maithani
Software Developer II
Helpshift Pvt Ltd
On Fri, Jan 17, 2020 at 10:37 AM Somya Maithani
wrote:
> Hey Team,
>
> *Problem*
> Recently, we were trying to upgrade Flink inf
Hey Team,
*Problem*
Recently, we were trying to upgrade Flink infrastructure to version 1.9.1
and we noticed that a week old offset was consumed from Kafka even though
the configuration says latest.
*Pretext*
1. Our current Flink version in production is 1.2.1.
2. We use RocksDB + Hadoop as our b
