Thanks for the explanation
LGTM
Thanks,
Penghui
On Tue, Aug 23, 2022 at 11:41 PM Dave Fisher wrote:
> While not a comment about this proposal I have a comment another split
> bundle concept.
>
> Manually split out a topic into its own bundle.
>
> Say a bundle is 0x to 0x0200 with 5
While not a comment about this proposal I have a comment another split bundle
concept.
Manually split out a topic into its own bundle.
Say a bundle is 0x to 0x0200 with 5 topics.
t1 at 0x0010
t2 at 0x0030
t3 at 0x0110
t4 at 0x0123
t5 at 0x01AE
Let’s split out
"It's a good idea to improve the bundle split for the case that the traffic
of the topic doesn't change drastically
Otherwise, we should not use this policy. or can we use it for all cases?"
1.It is suitable for scenarios where topic traffic is relatively stable. In
addition, we can also adjust
It's a good idea to improve the bundle split for the case that the traffic
of the topic doesn't change drastically
Otherwise, we should not use this policy. or can we use it for all cases?
1.It is suitable for scenarios where topic traffic is relatively stable. In
addition, we can also adjust
It's a good idea to improve the bundle split for the case that the traffic
of the topic doesn't change drastically
Otherwise, we should not use this policy. or can we use it for all cases?
I think It should be documented in the proposal.
I have some questions
- do we need to consider the
The implementation logic has been modified, and the corresponding example has
also been modified as follows:
## Motivation
As we all know, Bundle split has 3 algorithms:
- range_equally_divide
- topic_count_equally_divide
- specified_positions_divide
However, none of these algorithms can
IMHO a better solution would be to allow the bundles to be split into more
units in a single operation. Instead of splitting one bundle into 2, split it
into 4 or 8….. If we do that then we can make that configurable (with a
reasonable default).
On 2022/07/26 08:39:29 lordcheng10 wrote:
>
When the broker is just started, the traffic or QPS may not be stable at this
time.
At this time, we can solve the problem by adjusting the time interval of the
first load reporting task. At present, the time interval for executing the
report task is 5 seconds, and it is not configurable.
I can
Hi Pulsar Community,
This is a PIP discussion on how to support splitting bundles according to flow
or qps.
The issue can be found: https://github.com/apache/pulsar/issues/16782
I copy the content here for convenience, any suggestions are welcome and
appreciated.
## Motivation
As we
When the broker is just started, the traffic or QPS may not be stable at this
time.
At this time, we can solve the problem by adjusting the time interval of the
first load reporting task.
The time interval for executing the report task is 5 seconds, and it is not
configurable.
I can add a
1. Similar problems also exist in topic_count_equally_divide. In some
scenarios, topics are also temporary, because they will be created and
destroyed continuously.
2. The traffic and qps of most scenarios are relatively stable. For scenarios
where traffic and qps change greatly and frequently,
Hi,
This is an interesting proposal and sure it can solve quite some load balance
case.
My concern on this is that, relative speaking, the QPS and flow data is
temporary,
it may changes a lot, so it's easy to create some bad cases, especially on
startup.
Any idea on this matter?
Thanks,
The PIP number is duplicated, the updated PIP number is PIP-169.
Hi Pulsar Community,
This is a PIP discussion on how to support splitting bundles according to flow
or qps.
The issue can be found:https://github.com/apache/pulsar/issues/16782
I copy the content here for convenience, any suggestions are welcome and
appreciated.
## Motivation
As we all
14 matches
Mail list logo