On Sat, Jul 13, 2019, at 17:54, George Li wrote:
> I just take a look at the updated KIP-455 again. I noticed
> this "targetReplicas" is removed and instead, put "addingReplicas" &
> "removingReplicas". So where does the new reassigned "targetReplicas"
> is stored? When all replicas in ISR, and
On Mon, Jul 15, 2019, at 11:51, Jason Gustafson wrote:
> Hi Colin,
>
> A few more questions below:
>
> 1. The KIP says that the --zookeeper option will be removed from
> kafka-reassign.sh. Do you mean that it will be deprecated and eventually
> removed?
Keeping the --zookeeper flag here would be
On Mon, Jul 15, 2019, at 14:31, Stanislav Kozlovski wrote:
> Hey George,
>
> > Different replica threads for throttling
> The reason we can't support throttling for reassigning partitions right now
> is because we have no good way of telling whether a replica is part of an
> ongoing reassignment o
Hey George,
> Different replica threads for throttling
The reason we can't support throttling for reassigning partitions right now
is because we have no good way of telling whether a replica is part of an
ongoing reassignment or not.
KIP-455 now gives us this knowledge with the `addingReplicas` an
On Sat, Jul 13, 2019, at 17:54, George Li wrote:
> Hi Stanislav,
>
> sorry for the late reply. comments below:
>
> > Thanks for the reminder. A lot of your suggestions are outlined in the
> > "Future Work" section of KIP-455. The pointer towards different
> > ReplicaFetcher thread pools is inte
Hi Colin,
A few more questions below:
1. The KIP says that the --zookeeper option will be removed from
kafka-reassign.sh. Do you mean that it will be deprecated and eventually
removed?
2. The KIP mentions that AR is shrunk iteratively. Does this have a
benefit? The downside is that we lose track
Hi Stanislav,
sorry for the late reply. comments below:
> Thanks for the reminder. A lot of your suggestions are outlined in the
> "Future Work" section of KIP-455. The pointer towards different
> ReplicaFetcher thread pools is interesting -- do you think there's
> much value in that? My intuit
On 12.07.2019 12:05, Stanislav Kozlovski wrote:
> We will
> not support fetching the progress of any particular reassignment call from
> the server-side. If the client is interested in a particular reassignment's
> progress, it should know what that reassignment consisted of and query
> those par
Hey Jan,
We will support multiple Admin APIs that can fetch on-going assignments at
different granularities, regardless of which (JVM) process scheduled it.
If a process doesn't know which partitions/topics are being reassigned, it
can ask for all on-going reassignments in the Kafka cluster.
You w
Great KIP,
pure java cruise-control would be a nice thing to have <3
I just want to ask what the opinions are on a way to get the Futures of
the assignment back. Say accross JVms.
Best Jan
On 02.07.2019 19:47, Stanislav Kozlovski wrote:
> Hey there, I need to start a new thread on KIP-455. I th
Hello again,
I've further iterated on the design and I think that we can restart the
vote soon if there are no major comments.
I have again edited the KIP. Key changes are:
- Use the `topics/[topic]` znode to store desired reassignments. There were
thoughts that we could not update the replicaSet
Hey there everybody,
I've edited the KIP. Here is a diff of the changes -
https://cwiki.apache.org/confluence/pages/diffpagesbyversion.action?pageId=112820260&selectedPageVersions=13&selectedPageVersions=11
Specifically,
- new AdminAPI for listing ongoing reassignments for _a given list of topics_
> Now that we support multiple reassignment requests, users may add execute>
> them incrementally. Suppose something goes horribly wrong and they want to>
> revert as quickly as possible - they would need to run the tool with>
> multiple rollback JSONs. I think that it would be useful to ha
Hey Colin,
I've got A couple more minor questions:
AlterPartitionAssignmentsRequest: Should we make the `Topics` and
`Partitions` field nullable as well, so we can appropriately cancel
pending reassignments at all levels of granularity? I think this will
be consistent with the rest of the request
On Wed, Jul 3, 2019, at 11:22, Stanislav Kozlovski wrote:
> Hey Colin,
>
> Thanks for the detailed reply!
>
> > Not really. Cruise Control, for example, tries to update the znode even
> > when it's not empty. This doesn't work, of course. This is an existing
> > problem with the old API-- wh
Hey Colin,
Thanks for the detailed reply!
> Not really. Cruise Control, for example, tries to update the znode even when
> it's not empty. This doesn't work, of course. This is an existing problem
> with the old API-- which the new API fixes, of course.
I am not sure that's the case - I thi
On Tue, Jul 2, 2019, at 10:47, Stanislav Kozlovski wrote:
> Hey there, I need to start a new thread on KIP-455. I think there might be
> an issue with the mailing server. For some reason, my replies to the
> previous discussion thread could not be seen by others. After numerous
> attempts, Colin su
17 matches
Mail list logo
