Remove CompletedFetch type from Fetcher #7228
Conversation
|
Retest this please. |
seglo
force-pushed
the
seglo/KAFKA-8822
branch
2 times, most recently
from
d33d30f to
afd1f4c
Compare
hachikuji
left a comment
hachikuji
left a comment
There was a problem hiding this comment.
Thanks, nice cleanup! Left a few small comments.
| @@ -1422,8 +1418,9 @@ private void drain() { | |||
| } | |||
| } | |||
|
|
|||
| // TODO: this has no usages. remove? | |||
There was a problem hiding this comment.
Yeah, let's remove if it's unused. We can add back if we need it.
| @@ -1379,10 +1368,12 @@ public static Sensor throttleTimeSensor(Metrics metrics, FetcherMetricsRegistry | |||
|
|
|||
| private class PartitionRecords { | |||
There was a problem hiding this comment.
What do you think about taking the name CompletedFetch here. It seemed a little more descriptive.
There was a problem hiding this comment.
I agree, since it's private to the Fetcher anyway. I've renamed it and a few obvious related variables and javadocs. There are some other things that could be renamed too, but I wanted to see what you thought first.
CompletedFetch.isFetched- This sounds a little strange now.isFetched == truewhen the records are drained which normally happens after it's been fully consumed. We could rename it toisDrainedorisConsumed.Fetcher.nextInLineRecords- I assume this was derived from thePartitionRecordstype as well, but I'm not sure. It could be renamed tonextInLineCompletedFetchornextInLineFetch.
There was a problem hiding this comment.
Sorry, missed this comment. Both of your suggestions sound good to me. I like isConsumed and nextInLineFetch.
| FetchResponse.PartitionData<Records> partitionData, | ||
| FetchResponseMetricAggregator metricAggregator, | ||
| Iterator<? extends RecordBatch> batches, | ||
| Long fetchedOffset, |
There was a problem hiding this comment.
nit: fetchOffset instead of fetchedOffset? I think this is meant to describe the offset from the fetch request.
seglo
force-pushed
the
seglo/KAFKA-8822
branch
2 times, most recently
from
8cc2e67 to
097d0dc
Compare
|
retest this please |
|
@hachikuji Thanks. There were a couple more things that could be renamed if you want. See this comment I made earlier in the PR: #7228 (comment) If you have a moment could you take a quick look at my comments in the benchmark PR? #7221 |
|
@seglo Sorry, missed your comment. Your suggestions sound good to me. |
seglo
force-pushed
the
seglo/KAFKA-8822
branch
from
097d0dc to
b28cc84
Compare
|
Ok, I think it's ready to go. I'll watch to get a clean test run. |
|
Retest this please. |
2 participants
Jira issue KAFKA-8822.
The
CompletedFetchtype is no longer required due to the work merged in PR #6988 (KAFKA-7548). This PR consolidatesCompletedFetchintoPartitionRecordsas discussed in #6988 (comment) with @hachikuji.