Hi Dong,
Thanks for considering my suggestions.
Based on your comments, I realized that my suggestion was not complete with
regard to KafkaConsumer API vs. consumer-broker protocol. While I propose
to keep KafkaConsumer#seek() unchanged and take offset only, the underlying
consumer will send the next FetchRequest() to broker with offset and
leaderEpoch if it is known (based on leader epoch cache in consumer) — note
that this is different from the current KIP, which suggests to always send
unknown leader epoch after seek(). This way, if the consumer and a broker
agreed on the point of non-divergence, which is some {offset, leaderEpoch}
pair, the new leader which causes another truncation (even further back)
will be able to detect new divergence and restart the process of finding
the new point of non-divergence. So, to answer your question, If the
truncation happens just after the user calls
KafkaConsumer#findOffsets(offset, leaderEpoch) followed by seek(offset),
the user will not seek to the wrong position without knowing that
truncation has happened, because the consumer will get another truncation
error, and seek again.
I am afraid, I did not understand your second question. Let me summarize my
suggestions again, and then give an example to hopefully make my
suggestions more clear. Also, the last part of my example shows how the
use-case in your first question will work. If it does not answer your
second question, would you mind clarifying? I am also focusing on the case
of a consumer having enough entries in the cache. The case of restarting
from committed offset either stored externally or internally will probably
need to be discussed more.
Let me summarize my suggestion again:
1) KafkaConsumer#seek() and KafkaConsumer#position() remains unchanged
2) New KafkaConsumer#findOffsets() takes {offset, leaderEpoch} pair per
topic partition and returns offset per topic partition.
3) FetchRequest() to broker after KafkaConsumer#seek() will contain the
offset set by seek and leaderEpoch that corresponds to the offset based on
leader epoch cache in the consumer.
The rest of this e-mail is a long and contrived example with several log
truncations and unclean leader elections to illustrate the API and your
first use-case. Suppose we have three brokers. Initially, Broker A, B, and
C has one message at offset 0 with leader epoch 0. Then, Broker A goes down
for some time. Broker B becomes a leader with epoch 1, and writes messages
to offsets 1 and 2. Broker C fetches offset 1, but before fetching offset
2, becomes a leader with leader epoch 2 and writes a message at offset 2.
Here is the state of brokers at this point:
> Broker A:
> offset 0, epoch 0 <— leader
> goes down…
> Broker B:
> offset 0, epoch 0
> offset 1, epoch 1 <- leader
> offset 2, epoch 1
Broker C:
> offset 0, epoch 0
> offset 1, epoch 1
> offset 2, epoch 2 <— leader
Before Broker C becomes a leader with leader epoch 2, the consumer consumed
the following messages from broker A and broker B:
{offset=0, leaderEpoch=0}, {offset=1, leaderEpoch=1}, {offset=2,
leaderEpoch=1}.
Consumer’s leader epoch cache at this point contains the following entries:
(leaderEpoch=0, startOffset=0)
(leaderEpoch=1, startOffset=1)
endOffset = 3
Then, broker B becomes the follower of broker C, truncates and starts
fetching from offset 2.
Consumer sends fetchRequest(offset=3, leaderEpoch=1) and gets LOG_TRUNCATION
error from broker C.
In response, the client calls KafkaConsumer#findOffsets(offset=3,
leaderEpoch=1). The underlying consumer sends
OffsetsForLeaderEpoch(leaderEpoch=1), broker C responds with
{leaderEpoch=1, endOffset=2}. So, KafkaConsumer#findOffsets(offset=3,
leaderEpoch=1) returns offset=2.
In response, consumer calls KafkaConsumer@seek(offset=2) followed by
poll(), which results in FetchRequest(offset=2, leaderEpoch=1) to broker C.
I will continue with this example with the goal to answer your first
question about truncation just after findOffsets() followed by seek():
Suppose, brokers B and C go down, and broker A comes up and becomes a
leader with leader epoch 3, and writes a message to offset 1. Suppose, this
happens before the consumer gets response from broker C to the previous
fetch request: FetchRequest(offset=2, leaderEpoch=1).
Consumer re-sends FetchRequest(offset=2, leaderEpoch=1) to broker A, which
returns LOG_TRUNCATION error, because broker A has leader epoch 3 > leader
epoch in FetchRequest with starting offset = 1 < offset 2 in FetchRequest().
In response, the user calls KafkaConsumer#findOffsets(offset=2,
leaderEpoch=1). The underlying consumer sends
OffsetsForLeaderEpoch(leaderEpoch=1), broker A responds with
{leaderEpoch=0, endOffset=1}; the underlying consumer finds leaderEpoch = 0
in its cache with end offset == 1, which results in
KafkaConsumer#findOffsets(offset=2, leaderEpoch=1) returning offset = 1.
In response, the user calls KafkaConsumer@seek(offset=1) followed by
poll(), which results in FetchRequest(offset=1, leaderEpoch=0) to broker A,
which responds with message at offset 1, leader epoch 3.
I will think some more about consumers restarting from committed offsets,
and send a follow up.
Thanks,
Anna
On Sat, Jul 7, 2018 at 1:36 AM Dong Lin <lindon...@gmail.com> wrote:
> Hey Anna,
>
> Thanks much for the thoughtful reply. It makes sense to different between
> "seeking to a message" and "seeking to a position". I have to questions
> here:
>
> - For "seeking to a message" use-case, with the proposed approach user
> needs to call findOffset(offset, leaderEpoch) followed by seek(offset). If
> message truncation and message append happen immediately after
> findOffset(offset,
> leaderEpoch) but before seek(offset), it seems that user will seek to the
> wrong message without knowing the truncation has happened. Would this be a
> problem?
>
> - For "seeking to a position" use-case, it seems that there can be two
> positions, i.e. earliest and latest. So these two cases can be
> Consumer.fulfilled by seekToBeginning() and Consumer.seekToEnd(). Then it
> seems that user will only need to call position() and seek() for "seeking
> to a message" use-case?
>
> Thanks,
> Dong
>
>
> On Wed, Jul 4, 2018 at 12:33 PM, Anna Povzner <a...@confluent.io> wrote:
>
> > Hi Jason and Dong,
> >
> >
> > I’ve been thinking about your suggestions and discussion regarding
> > position(), seek(), and new proposed API.
> >
> >
> > Here is my thought process why we should keep position() and seek() API
> > unchanged.
> >
> >
> > I think we should separate {offset, leader epoch} that uniquely
> identifies
> > a message from an offset that is a position. In some cases, offsets
> > returned from position() could be actual consumed messages by this
> consumer
> > identified by {offset, leader epoch}. In other cases, position() returns
> > offset that was not actually consumed. Suppose, the user calls position()
> > for the last offset. Suppose we return {offset, leader epoch} of the
> > message currently in the log. Then, the message gets truncated before
> > consumer’s first poll(). It does not make sense for poll() to fail in
> this
> > case, because the log truncation did not actually happen from the
> consumer
> > perspective. On the other hand, as the KIP proposes, it makes sense for
> the
> > committed() method to return {offset, leader epoch} because those offsets
> > represent actual consumed messages.
> >
> >
> > The same argument applies to the seek() method — we are not seeking to a
> > message, we are seeking to a position.
> >
> >
> > I like the proposal to add KafkaConsumer#findOffsets() API. I am assuming
> > something like:
> >
> > Map<TopicPartition, Long> findOffsets(Map<TopicPartition, OffsetAndEpoch>
> > offsetsToSearch)
> >
> > Similar to seek() and position(), I think findOffsets() should return
> > offset without leader epoch, because what we want is the offset that we
> > think is closest to the not divergent message from the given consumed
> > message. Until the consumer actually fetches the message, we should not
> let
> > the consumer store the leader epoch for a message it did not consume.
> >
> >
> > So, the workflow will be:
> >
> > 1) The user gets LogTruncationException with {offset, leader epoch of the
> > previous message} (whatever we send with new FetchRecords request).
> >
> > 2) offset = findOffsets(tp -> {offset, leader epoch})
> >
> > 3) seek(offset)
> >
> >
> > For the use-case where the users store committed offsets externally:
> >
> > 1) Such users would have to track the leader epoch together with an
> offset.
> > Otherwise, there is no way to detect later what leader epoch was
> associated
> > with the message. I think it’s reasonable to ask that from users if they
> > want to detect log truncation. Otherwise, they will get the current
> > behavior.
> >
> >
> > If the users currently get an offset to be stored using position(), I see
> > two possibilities. First, they call save offset returned from position()
> > that they call before poll(). In that case, it would not be correct to
> > store {offset, leader epoch} if we would have changed position() to
> return
> > {offset, leader epoch} since actual fetched message could be different
> > (from the example I described earlier). So, it would be more correct to
> > call position() after poll(). However, the user already gets
> > ConsumerRecords at this point, from which the user can extract {offset,
> > leader epoch} of the last message.
> >
> >
> > So, I like the idea of adding a helper method to ConsumerRecords, as
> Jason
> > proposed, something like:
> >
> > public OffsetAndEpoch lastOffsetWithLeaderEpoch(), where OffsetAndEpoch
> is
> > a data struct holding {offset, leader epoch}.
> >
> >
> > In this case, we would advise the user to follow the workflow: poll(),
> get
> > {offset, leader epoch} from ConsumerRecords#lastOffsetWithLeaderEpoch(),
> > save offset and leader epoch, process records.
> >
> >
> > 2) When the user needs to seek to the last committed offset, they call
> new
> > findOffsets(saved offset, leader epoch), and then seek(offset).
> >
> >
> > What do you think?
> >
> >
> > Thanks,
> >
> > Anna
> >
> >
> > On Tue, Jul 3, 2018 at 4:06 PM Dong Lin <lindon...@gmail.com> wrote:
> >
> > > Hey Jason,
> > >
> > > Thanks much for your thoughtful explanation.
> > >
> > > Yes the solution using findOffsets(offset, leaderEpoch) also works. The
> > > advantage of this solution it adds only one API instead of two APIs.
> The
> > > concern is that its usage seems a bit more clumsy for advanced users.
> > More
> > > specifically, advanced users who store offsets externally will always
> > need
> > > to call findOffsets() before calling seek(offset) during consumer
> > > initialization. And those advanced users will need to manually keep
> track
> > > of the leaderEpoch of the last ConsumerRecord.
> > >
> > > The other solution may be more user-friendly for advanced users is to
> add
> > > two APIs, `void seek(offset, leaderEpoch)` and `(offset, epoch) =
> > > offsetEpochs(topicPartition)`.
> > >
> > > I kind of prefer the second solution because it is easier to use for
> > > advanced users. If we need to expose leaderEpoch anyway to safely
> > identify
> > > a message, it may be conceptually simpler to expose it directly in
> > > seek(...) rather than requiring one more translation using
> > > findOffsets(...). But I am also OK with the first solution if other
> > > developers also favor that one :)
> > >
> > > Thanks,
> > > Dong
> > >
> > >
> > > On Mon, Jul 2, 2018 at 11:10 AM, Jason Gustafson <ja...@confluent.io>
> > > wrote:
> > >
> > > > Hi Dong,
> > > >
> > > > Thanks, I've been thinking about your suggestions a bit. It is
> > > challenging
> > > > to make this work given the current APIs. One of the difficulties is
> > that
> > > > we don't have an API to find the leader epoch for a given offset at
> the
> > > > moment. So if the user does a seek to offset 5, then we'll need a new
> > API
> > > > to find the corresponding epoch in order to fulfill the new
> position()
> > > API.
> > > > Potentially we could modify ListOffsets to enable finding the leader
> > > epoch,
> > > > but I am not sure it is worthwhile. Perhaps it is reasonable for
> > advanced
> > > > usage to expect that the epoch information, if needed, will be
> > extracted
> > > > from the records directly? It might make sense to expose a helper in
> > > > `ConsumerRecords` to make this a little easier though.
> > > >
> > > > Alternatively, if we think it is important to have this information
> > > exposed
> > > > directly, we could create batch APIs to solve the naming problem. For
> > > > example:
> > > >
> > > > Map<TopicPartition, OffsetAndEpoch> positions();
> > > > void seek(Map<TopicPartition, OffsetAndEpoch> positions);
> > > >
> > > > However, I'm actually leaning toward leaving the seek() and
> position()
> > > APIs
> > > > unchanged. Instead, we can add a new API to search for offset by
> > > timestamp
> > > > or by offset/leader epoch. Let's say we call it `findOffsets`. If the
> > > user
> > > > hits a log truncation error, they can use this API to find the
> closest
> > > > offset and then do a seek(). At the same time, we deprecate the
> > > > `offsetsForTimes` APIs. We now have two use cases which require
> finding
> > > > offsets, so I think we should make this API general and leave the
> door
> > > open
> > > > for future extensions.
> > > >
> > > > By the way, I'm unclear about the desire to move part of this
> > > functionality
> > > > to AdminClient. Guozhang suggested this previously, but I think it
> only
> > > > makes sense for cross-cutting capabilities such as topic creation. If
> > we
> > > > have an API which is primarily useful by consumers, then I think
> that's
> > > > where it should be exposed. The AdminClient also has its own API
> > > integrity
> > > > and should not become a dumping ground for advanced use cases. I'll
> > > update
> > > > the KIP with the `findOffsets` API suggested above and we can see if
> > it
> > > > does a good enough job of keeping the API simple for common cases.
> > > >
> > > > Thanks,
> > > > Jason
> > > >
> > > >
> > > > On Sat, Jun 30, 2018 at 4:39 AM, Dong Lin <lindon...@gmail.com>
> wrote:
> > > >
> > > > > Hey Jason,
> > > > >
> > > > > Regarding seek(...), it seems that we want an API for user to
> > > initialize
> > > > > consumer with (offset, leaderEpoch) and that API should allow
> > throwing
> > > > > PartitionTruncationException. Suppose we agree on this, then
> > > > > seekToNearest() is not sufficient because it will always swallow
> > > > > PartitionTruncationException. Here we have two options. The first
> > > option
> > > > is
> > > > > to add API offsetsForLeaderEpochs() to translate (leaderEpoch,
> > offset)
> > > to
> > > > > offset. The second option is to have add seek(offset, leaderEpoch).
> > It
> > > > > seems that second option may be more simpler because it makes it
> > clear
> > > > that
> > > > > (offset, leaderEpoch) will be used to identify consumer's position
> > in a
> > > > > partition. And user only needs to handle
> PartitionTruncationException
> > > > from
> > > > > the poll(). In comparison the first option seems a bit harder to
> use
> > > > > because user have to also handle the PartitionTruncationException
> if
> > > > > offsetsForLeaderEpochs() returns different offset from
> user-provided
> > > > > offset. What do you think?
> > > > >
> > > > > If we decide to add API seek(offset, leaderEpoch), then we can
> decide
> > > > > whether and how to add API to translate (offset, leaderEpoch) to
> > > offset.
> > > > It
> > > > > seems that this API will be needed by advanced user to don't want
> > auto
> > > > > offset reset (so that it can be notified) but still wants to reset
> > > offset
> > > > > to closest. For those users if probably makes sense to only have
> the
> > > API
> > > > in
> > > > > AdminClient. offsetsForTimes() seems like a common API that will be
> > > > needed
> > > > > by user's of consumer in general, so it may be more reasonable to
> > stay
> > > in
> > > > > the consumer API. I don't have a strong opinion on whether
> > > > > offsetsForTimes() should be replaced by API in AdminClient.
> > > > >
> > > > > Though (offset, leaderEpoch) is needed to uniquely identify a
> message
> > > in
> > > > > general, it is only needed for advanced users who has turned on
> > unclean
> > > > > leader election, need to use seek(..), and don't want auto offset
> > > reset.
> > > > > Most other users probably just want to enable auto offset reset and
> > > store
> > > > > offset in Kafka. Thus we might want to keep the existing
> offset-only
> > > APIs
> > > > > (e.g. seek() and position()) for most users while adding new APIs
> for
> > > > > advanced users. And yes, it seems that we need new name for
> > position().
> > > > >
> > > > > Though I think we need new APIs to carry the new information (e.g.
> > > > > leaderEpoch), I am not very sure how that should look like. One
> > > possible
> > > > > option is those APIs in KIP-232. Another option is something like
> > this:
> > > > >
> > > > > `````
> > > > > class OffsetEpochs {
> > > > > long offset;
> > > > > int leaderEpoch;
> > > > > int partitionEpoch; // This may be needed later as discussed in
> > > > KIP-232
> > > > > ... // Hopefully these are all we need to identify message in
> > Kafka.
> > > > But
> > > > > if we need more then we can add new fields in this class.
> > > > > }
> > > > >
> > > > > OffsetEpochs offsetEpochs(TopicPartition);
> > > > >
> > > > > void seek(TopicPartition, OffsetEpochs);
> > > > > ``````
> > > > >
> > > > >
> > > > > Thanks,
> > > > > Dong
> > > > >
> > > > >
> > > > > On Fri, Jun 29, 2018 at 11:13 AM, Jason Gustafson <
> > ja...@confluent.io>
> > > > > wrote:
> > > > >
> > > > > > Hey Dong,
> > > > > >
> > > > > > Thanks for the feedback. The first three points are easy:
> > > > > >
> > > > > > 1. Yes, we should be consistent.
> > > > > > 2. Yes, I will add this.
> > > > > > 3. Yes, I think we should document the changes to the committed
> > > offset
> > > > > > schema. I meant to do this, but it slipped my mind.
> > > > > >
> > > > > > The latter questions are tougher. One option I was considering is
> > to
> > > > have
> > > > > > only `offsetsForLeaderEpochs` exposed from the consumer and to
> drop
> > > the
> > > > > new
> > > > > > seek() API. That seems more consistent with the current use of
> > > > > > `offsetsForTimes` (we don't have a separate `seekToTimestamp`
> API).
> > > An
> > > > > > alternative might be to take a page from the AdminClient API and
> > add
> > > a
> > > > > new
> > > > > > method to generalize offset lookup. For example, we could have
> > > > > > `lookupOffsets(LookupOptions)`. We could then deprecate
> > > > > `offsetsForTimes`
> > > > > > and this would open the door for future extensions without
> needing
> > > new
> > > > > > APIs.
> > > > > >
> > > > > > The case of position() is a little more annoying. It would have
> > been
> > > > > better
> > > > > > had we let this return an object so that it is easier to extend.
> > This
> > > > is
> > > > > > the only reason I didn't add the API to the KIP. Maybe we should
> > bite
> > > > the
> > > > > > bullet and fix this now? Unfortunately we'll have to come up
> with a
> > > new
> > > > > > name. Maybe `currentPosition`?
> > > > > >
> > > > > > Thoughts?
> > > > > >
> > > > > > -Jason
> > > > > >
> > > > > >
> > > > > > On Fri, Jun 29, 2018 at 10:40 AM, Dong Lin <lindon...@gmail.com>
> > > > wrote:
> > > > > >
> > > > > > > Regarding points 4) and 5) above, motivation for the
> alternative
> > > APIs
> > > > > is
> > > > > > > that, if we decide that leaderEpoch is equally important as
> > offset
> > > in
> > > > > > > identifying a message, then it may be reasonable to always
> > specify
> > > it
> > > > > > > wherever offset is currently required in the consumer API to
> > > > identify a
> > > > > > > message, e.g. position(), seek(). For example, since we allow
> > user
> > > to
> > > > > > > retrieve offset using position() instead of asking user to keep
> > > track
> > > > > of
> > > > > > > the offset of the latest ConsumerRecord, may be it will be more
> > > > > > consistent
> > > > > > > for user to also retrieve leaderEpoch using position()?
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > On Fri, Jun 29, 2018 at 10:30 AM, Dong Lin <
> lindon...@gmail.com>
> > > > > wrote:
> > > > > > >
> > > > > > > > Hey Jason,
> > > > > > > >
> > > > > > > > Thanks for the update. It looks pretty good. Just some minor
> > > > comments
> > > > > > > > below:
> > > > > > > >
> > > > > > > > 1) The KIP adds new error code "LOG_TRUNCATION" and new
> > exception
> > > > > > > TruncatedPartitionException.
> > > > > > > > Can we make the name more consistent, e.g.
> > > LogTruncationException?
> > > > > > > >
> > > > > > > > 2) Do we need to add UnknownLeaderEpochException as part of
> API
> > > > > change?
> > > > > > > >
> > > > > > > > 3) Not sure if the offset topic schema is also public API. If
> > so,
> > > > > maybe
> > > > > > > we
> > > > > > > > should also include the schema change in the API?
> > > > > > > >
> > > > > > > > 4) For users who store offset externally, currently they get
> > > offset
> > > > > > using
> > > > > > > > position(..), store the offset externally, and use seek(..)
> to
> > > > > > initialize
> > > > > > > > the consumer next time. After this KIP they will need to
> store
> > > and
> > > > > use
> > > > > > > the
> > > > > > > > leaderEpoch together with the offset. Should we also update
> the
> > > API
> > > > > so
> > > > > > > that
> > > > > > > > user can also get leaderEpoch from position(...)? Not sure if
> > it
> > > is
> > > > > OK
> > > > > > to
> > > > > > > > ask user to track the latest leaderEpoch of ConsumerRecord by
> > > > > > themselves.
> > > > > > > >
> > > > > > > > 5) Also for users who store offset externally, they need to
> > call
> > > > > > seek(..)
> > > > > > > > with leaderEpoch to initialize consumer. With current KIP
> users
> > > > need
> > > > > to
> > > > > > > > call seekToNearest(), whose name suggests that the final
> > position
> > > > may
> > > > > > be
> > > > > > > > different from what was requested. However, if users may want
> > to
> > > > > avoid
> > > > > > > auto
> > > > > > > > offset reset and be notified explicitly when there is log
> > > > truncation,
> > > > > > > then seekToNearest()
> > > > > > > > probably does not help here. Would it make sense to replace
> > > > > > > seekToNearest()
> > > > > > > > with seek(offset, leaderEpoch) + AminClient.
> > > > > > offsetsForLeaderEpochs(...)?
> > > > > > > >
> > > > > > > >
> > > > > > > > Thanks,
> > > > > > > > Dong
> > > > > > > >
> > > > > > > >
> > > > > > > > On Wed, Jun 27, 2018 at 3:57 PM, Jason Gustafson <
> > > > ja...@confluent.io
> > > > > >
> > > > > > > > wrote:
> > > > > > > >
> > > > > > > >> Hey Guozhang,
> > > > > > > >>
> > > > > > > >> That's fair. In fact, perhaps we do not need this API at
> all.
> > We
> > > > > > already
> > > > > > > >> have the new seek() in this KIP which can do the lookup
> based
> > on
> > > > > epoch
> > > > > > > for
> > > > > > > >> this use case. I guess we should probably call it
> > > seekToNearest()
> > > > > > though
> > > > > > > >> to
> > > > > > > >> make it clear that the final position may be different from
> > what
> > > > was
> > > > > > > >> requested.
> > > > > > > >>
> > > > > > > >> Thanks,
> > > > > > > >> Jason
> > > > > > > >>
> > > > > > > >> On Wed, Jun 27, 2018 at 3:20 PM, Guozhang Wang <
> > > > wangg...@gmail.com>
> > > > > > > >> wrote:
> > > > > > > >>
> > > > > > > >> > Hi Jason,
> > > > > > > >> >
> > > > > > > >> > I think it is less worthwhile to add
> > > > > KafkaConsumer#offsetsForLeader
> > > > > > > >> Epochs,
> > > > > > > >> > since probably only very advanced users are aware of the
> > > > > > leaderEpoch,
> > > > > > > >> and
> > > > > > > >> > hence ever care to use it anyways. It is more like an
> admin
> > > > client
> > > > > > > >> > operation than a consumer client operation: if the
> > motivation
> > > is
> > > > > to
> > > > > > > >> > facility customized reset policy, maybe adding it as
> > > > > > > >> > AdminClient#offsetsForLeaderEpochs
> > > > > > > >> > is better as it is not an aggressive assumption that for
> > such
> > > > > > advanced
> > > > > > > >> > users they are willing to use some admin client to get
> > further
> > > > > > > >> information?
> > > > > > > >> >
> > > > > > > >> >
> > > > > > > >> > Guozhang
> > > > > > > >> >
> > > > > > > >> >
> > > > > > > >> > On Wed, Jun 27, 2018 at 2:07 PM, Jason Gustafson <
> > > > > > ja...@confluent.io>
> > > > > > > >> > wrote:
> > > > > > > >> >
> > > > > > > >> > > Thanks for the feedback. I've updated the KIP.
> > Specifically
> > > I
> > > > > > > removed
> > > > > > > >> the
> > > > > > > >> > > "closest" reset option and the proposal to reset by
> > > timestamp
> > > > > when
> > > > > > > the
> > > > > > > >> > > precise truncation point cannot be determined. Instead,
> I
> > > > > proposed
> > > > > > > >> that
> > > > > > > >> > we
> > > > > > > >> > > always reset using the nearest epoch when a reset policy
> > is
> > > > > > defined
> > > > > > > >> > (either
> > > > > > > >> > > "earliest" or "latest"). Does that sound reasonable?
> > > > > > > >> > >
> > > > > > > >> > > One thing I am still debating is whether it would be
> > better
> > > to
> > > > > > have
> > > > > > > a
> > > > > > > >> > > separate API to find the closest offset using the leader
> > > > epoch.
> > > > > In
> > > > > > > the
> > > > > > > >> > > current KIP, I suggested to piggyback this information
> on
> > an
> > > > > > > >> exception,
> > > > > > > >> > but
> > > > > > > >> > > I'm beginning to think it would be better not to hide
> the
> > > > > lookup.
> > > > > > It
> > > > > > > >> is
> > > > > > > >> > > awkward to implement since it means delaying the
> exception
> > > and
> > > > > the
> > > > > > > API
> > > > > > > >> > may
> > > > > > > >> > > actually be useful when customizing reset logic if no
> auto
> > > > reset
> > > > > > > >> policy
> > > > > > > >> > is
> > > > > > > >> > > defined. I was thinking we can add an API like the
> > > following:
> > > > > > > >> > >
> > > > > > > >> > > Map<TopicPartition, OffsetAndEpoch>
> > > > > > > >> > > offsetsForLeaderEpochs(Map<TopicPartition, Integer>
> > > > > > epochsToSearch)
> > > > > > > >> > >
> > > > > > > >> > > Thoughts?
> > > > > > > >> > >
> > > > > > > >> > > -Jason
> > > > > > > >> > >
> > > > > > > >> > >
> > > > > > > >> > >
> > > > > > > >> > >
> > > > > > > >> > > On Wed, Jun 27, 2018 at 11:12 AM, Jason Gustafson <
> > > > > > > ja...@confluent.io
> > > > > > > >> >
> > > > > > > >> > > wrote:
> > > > > > > >> > >
> > > > > > > >> > > > @Dong
> > > > > > > >> > > >
> > > > > > > >> > > > Those are fair points. Both approaches require some
> > > > fuzziness
> > > > > to
> > > > > > > >> reset
> > > > > > > >> > > the
> > > > > > > >> > > > offset in these pathological scenarios and we cannot
> > > > guarantee
> > > > > > > >> > > > at-least-once delivery either way unless we have the
> > full
> > > > > > history
> > > > > > > of
> > > > > > > >> > > leader
> > > > > > > >> > > > epochs that were consumed. The KIP-101 logic may
> > actually
> > > be
> > > > > > more
> > > > > > > >> > > accurate
> > > > > > > >> > > > than using timestamps because it does not depend on
> the
> > > > > messages
> > > > > > > >> which
> > > > > > > >> > > are
> > > > > > > >> > > > written after the unclean leader election. The case
> > we're
> > > > > > talking
> > > > > > > >> about
> > > > > > > >> > > > should be extremely rare in practice anyway. I also
> > agree
> > > > that
> > > > > > we
> > > > > > > >> may
> > > > > > > >> > not
> > > > > > > >> > > > want to add new machinery if it only helps the old
> > message
> > > > > > format.
> > > > > > > >> Ok,
> > > > > > > >> > > > let's go ahead and drop the timestamp.
> > > > > > > >> > > >
> > > > > > > >> > > > @Guozhang
> > > > > > > >> > > >
> > > > > > > >> > > > * My current understanding is that, with unclean
> leader
> > > > > election
> > > > > > > >> turned
> > > > > > > >> > > on,
> > > > > > > >> > > >> exactly-once is out of the window since we cannot
> > > guarantee
> > > > > > that
> > > > > > > >> all
> > > > > > > >> > > >> committed message markers will not be lost. And hence
> > > there
> > > > > is
> > > > > > no
> > > > > > > >> need
> > > > > > > >> > > to
> > > > > > > >> > > >> have special handling logic for LOG_TRUNCATED or OOR
> > > error
> > > > > > codes
> > > > > > > >> with
> > > > > > > >> > > >> read.committed turned on. Is that right?
> > > > > > > >> > > >
> > > > > > > >> > > >
> > > > > > > >> > > > Yes, that's right. EoS and unclean leader election
> don't
> > > mix
> > > > > > well.
> > > > > > > >> It
> > > > > > > >> > may
> > > > > > > >> > > > be worth considering separately whether we should try
> to
> > > > > > reconcile
> > > > > > > >> the
> > > > > > > >> > > > transaction log following an unclean leader election.
> At
> > > > least
> > > > > > we
> > > > > > > >> may
> > > > > > > >> > be
> > > > > > > >> > > > able to prevent dangling transactions from blocking
> > > > consumers.
> > > > > > > This
> > > > > > > >> KIP
> > > > > > > >> > > > does not address this problem.
> > > > > > > >> > > >
> > > > > > > >> > > > * MINOR: "if the epoch is greater than the minimum
> > > expected
> > > > > > epoch,
> > > > > > > >> that
> > > > > > > >> > > the
> > > > > > > >> > > >> new epoch does not begin at an earlier offset than
> the
> > > > fetch
> > > > > > > >> offset.
> > > > > > > >> > In
> > > > > > > >> > > >> the latter case, the leader can respond with a new
> > > > > > LOG_TRUNCATION
> > > > > > > >> > error
> > > > > > > >> > > >> code" should it be "does not begin at a later offset
> > than
> > > > the
> > > > > > > fetch
> > > > > > > >> > > >> offset"?
> > > > > > > >> > > >
> > > > > > > >> > > >
> > > > > > > >> > > > I think the comment is correct, though the phrasing
> may
> > be
> > > > > > > >> confusing.
> > > > > > > >> > We
> > > > > > > >> > > > know truncation has occurred if there exists a larger
> > > epoch
> > > > > > with a
> > > > > > > >> > > starting
> > > > > > > >> > > > offset that is lower than the fetch offset. Let me try
> > to
> > > > > > rephrase
> > > > > > > >> > this.
> > > > > > > >> > > >
> > > > > > > >> > > > Thanks,
> > > > > > > >> > > > Jason
> > > > > > > >> > > >
> > > > > > > >> > > > On Wed, Jun 27, 2018 at 9:17 AM, Guozhang Wang <
> > > > > > > wangg...@gmail.com>
> > > > > > > >> > > wrote:
> > > > > > > >> > > >
> > > > > > > >> > > >> Jason, thanks for the KIP. A few comments:
> > > > > > > >> > > >>
> > > > > > > >> > > >> * I think Dong's question about whether to use
> > > > > timestamp-based
> > > > > > > >> > approach
> > > > > > > >> > > >> v.s. start-offset-of-first-larger-epoch is valid;
> more
> > > > > > > >> specifically,
> > > > > > > >> > > with
> > > > > > > >> > > >> timestamp-based approach we may still be reseting to
> an
> > > > > offset
> > > > > > > >> falling
> > > > > > > >> > > >> into
> > > > > > > >> > > >> the truncated interval, and hence we may still miss
> > some
> > > > > data,
> > > > > > > i.e.
> > > > > > > >> > not
> > > > > > > >> > > >> guaranteeing at-least-once still. With the
> > > > > > > >> > > >> start-offset-of-first-larger-epoch, I'm not sure if
> it
> > > > will
> > > > > > > >> guarantee
> > > > > > > >> > > no
> > > > > > > >> > > >> valid data is missed when we have consecutive log
> > > > truncations
> > > > > > > >> (maybe
> > > > > > > >> > we
> > > > > > > >> > > >> need to look back into details of KIP-101 to figure
> it
> > > > out).
> > > > > If
> > > > > > > the
> > > > > > > >> > > latter
> > > > > > > >> > > >> can indeed guarantee at least once, we could consider
> > > using
> > > > > > that
> > > > > > > >> > > approach.
> > > > > > > >> > > >>
> > > > > > > >> > > >> * My current understanding is that, with unclean
> leader
> > > > > > election
> > > > > > > >> > turned
> > > > > > > >> > > >> on,
> > > > > > > >> > > >> exactly-once is out of the window since we cannot
> > > guarantee
> > > > > > that
> > > > > > > >> all
> > > > > > > >> > > >> committed message markers will not be lost. And hence
> > > there
> > > > > is
> > > > > > no
> > > > > > > >> need
> > > > > > > >> > > to
> > > > > > > >> > > >> have special handling logic for LOG_TRUNCATED or OOR
> > > error
> > > > > > codes
> > > > > > > >> with
> > > > > > > >> > > >> read.committed turned on. Is that right?
> > > > > > > >> > > >>
> > > > > > > >> > > >> * MINOR: "if the epoch is greater than the minimum
> > > expected
> > > > > > > epoch,
> > > > > > > >> > that
> > > > > > > >> > > >> the
> > > > > > > >> > > >> new epoch does not begin at an earlier offset than
> the
> > > > fetch
> > > > > > > >> offset.
> > > > > > > >> > In
> > > > > > > >> > > >> the latter case, the leader can respond with a new
> > > > > > LOG_TRUNCATION
> > > > > > > >> > error
> > > > > > > >> > > >> code" should it be "does not begin at a later offset
> > than
> > > > the
> > > > > > > fetch
> > > > > > > >> > > >> offset"?
> > > > > > > >> > > >>
> > > > > > > >> > > >>
> > > > > > > >> > > >>
> > > > > > > >> > > >> Guozhang
> > > > > > > >> > > >>
> > > > > > > >> > > >>
> > > > > > > >> > > >> On Tue, Jun 26, 2018 at 6:51 PM, Dong Lin <
> > > > > lindon...@gmail.com
> > > > > > >
> > > > > > > >> > wrote:
> > > > > > > >> > > >>
> > > > > > > >> > > >> > Hey Jason,
> > > > > > > >> > > >> >
> > > > > > > >> > > >> > Thanks for the explanation.
> > > > > > > >> > > >> >
> > > > > > > >> > > >> > Please correct me if this is wrong. The "unknown
> > > > truncation
> > > > > > > >> offset"
> > > > > > > >> > > >> > scenario happens when consumer does not have the
> full
> > > > > > > >> leaderEpoch ->
> > > > > > > >> > > >> offset
> > > > > > > >> > > >> > mapping. In this case we can still use the
> > > KIP-101-based
> > > > > > > >> approach to
> > > > > > > >> > > >> > truncate offset to "start offset of the first
> Leader
> > > > Epoch
> > > > > > > larger
> > > > > > > >> > than
> > > > > > > >> > > >> last
> > > > > > > >> > > >> > epoch of the consumer" but it may be inaccurate. So
> > the
> > > > KIP
> > > > > > > >> chooses
> > > > > > > >> > to
> > > > > > > >> > > >> use
> > > > > > > >> > > >> > the timestamp-based approach which is also
> > best-effort.
> > > > > > > >> > > >> >
> > > > > > > >> > > >> > If this understanding is correct, for "closest"
> > offset
> > > > > reset
> > > > > > > >> policy
> > > > > > > >> > > and
> > > > > > > >> > > >> > "unknown truncation offset" scenario, I am
> wondering
> > > > > whether
> > > > > > it
> > > > > > > >> > maybe
> > > > > > > >> > > >> > better to replace timestamp-based approach with
> > KIP-101
> > > > > based
> > > > > > > >> > > approach.
> > > > > > > >> > > >> In
> > > > > > > >> > > >> > comparison to timestamp-based approach, the
> > > KIP-101-based
> > > > > > > >> approach
> > > > > > > >> > > >> seems to
> > > > > > > >> > > >> > simplify the API a bit since user does not need to
> > > > > understand
> > > > > > > >> > > timestamp.
> > > > > > > >> > > >> > Similar to the timestamp-based approach, both
> > > approaches
> > > > > are
> > > > > > > >> > > best-effort
> > > > > > > >> > > >> > and do not guarantee that consumer can consume all
> > > > > messages.
> > > > > > It
> > > > > > > >> is
> > > > > > > >> > not
> > > > > > > >> > > >> like
> > > > > > > >> > > >> > KIP-279 which guarantees that follower broker can
> > > consume
> > > > > all
> > > > > > > >> > messages
> > > > > > > >> > > >> from
> > > > > > > >> > > >> > the leader.
> > > > > > > >> > > >> >
> > > > > > > >> > > >> > Then it seems that the remaining difference is
> mostly
> > > > about
> > > > > > > >> > accuracy,
> > > > > > > >> > > >> i.e.
> > > > > > > >> > > >> > how much message will be duplicated or missed in
> the
> > > > > "unknown
> > > > > > > >> > > truncation
> > > > > > > >> > > >> > offset" scenario. Not sure either one is clearly
> > better
> > > > > than
> > > > > > > the
> > > > > > > >> > > other.
> > > > > > > >> > > >> > Note that there are two scenarios mentioned in
> > KIP-279
> > > > > which
> > > > > > > are
> > > > > > > >> not
> > > > > > > >> > > >> > addressed by KIP-101. Both scenarios require quick
> > > > > leadership
> > > > > > > >> change
> > > > > > > >> > > >> > between brokers, which seems to suggest that the
> > offset
> > > > > based
> > > > > > > >> > obtained
> > > > > > > >> > > >> > by "start
> > > > > > > >> > > >> > offset of the first Leader Epoch larger than last
> > epoch
> > > > of
> > > > > > the
> > > > > > > >> > > consumer"
> > > > > > > >> > > >> > under these two scenarios may be very close to the
> > > offset
> > > > > > > >> obtained
> > > > > > > >> > by
> > > > > > > >> > > >> the
> > > > > > > >> > > >> > message timestamp. Does this sound reasonable?
> > > > > > > >> > > >> >
> > > > > > > >> > > >> > Good point that users on v1 format can get benefit
> > with
> > > > > > > timestamp
> > > > > > > >> > > based
> > > > > > > >> > > >> > approach. On the other hand it seems like a short
> > term
> > > > > > benefit
> > > > > > > >> for
> > > > > > > >> > > users
> > > > > > > >> > > >> > who have not migrated. I am just not sure whether
> it
> > is
> > > > > more
> > > > > > > >> > important
> > > > > > > >> > > >> than
> > > > > > > >> > > >> > designing a better API.
> > > > > > > >> > > >> >
> > > > > > > >> > > >> > Also, for both "latest" and "earliest" reset
> policy,
> > do
> > > > you
> > > > > > > >> think it
> > > > > > > >> > > >> would
> > > > > > > >> > > >> > make sense to also use the KIP-101 based approach
> to
> > > > > truncate
> > > > > > > >> offset
> > > > > > > >> > > for
> > > > > > > >> > > >> > the "unknown truncation offset" scenario?
> > > > > > > >> > > >> >
> > > > > > > >> > > >> >
> > > > > > > >> > > >> > Thanks,
> > > > > > > >> > > >> > Dong
> > > > > > > >> > > >> >
> > > > > > > >> > > >>
> > > > > > > >> > > >>
> > > > > > > >> > > >>
> > > > > > > >> > > >> --
> > > > > > > >> > > >> -- Guozhang
> > > > > > > >> > > >>
> > > > > > > >> > > >
> > > > > > > >> > > >
> > > > > > > >> > >
> > > > > > > >> >
> > > > > > > >> >
> > > > > > > >> >
> > > > > > > >> > --
> > > > > > > >> > -- Guozhang
> > > > > > > >> >
> > > > > > > >>
> > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
