[
https://issues.apache.org/jira/browse/KUDU-613?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17912688#comment-17912688
]
ASF subversion and git services commented on KUDU-613:
------------------------------------------------------
Commit 622683201dc1b5c1b056d84565ee6cae8964ddd2 in kudu's branch
refs/heads/master from Mahesh Reddy
[ https://gitbox.apache.org/repos/asf?p=kudu.git;h=622683201 ]
KUDU-613: Make segment metrics more relevant
This patch removes the segment-level lookup metrics
along with the upgrade and downgrade counter metrics.
It instead adds histogram metrics to keep track of
the stats of upgrades and downgrades per entry. The
total number of upgrades and downgrades for the cache
can be derived from these new metrics using the field
'total_count'.
Counters to keep track of upgrades and downgrades
are also added on the handle level.
Change-Id: Id44828be41a19593e6808debf7e9ba8e1fc4dcca
Reviewed-on: http://gerrit.cloudera.org:8080/22253
Tested-by: Alexey Serbin <ale...@apache.org>
Reviewed-by: Alexey Serbin <ale...@apache.org>
> Scan-resistant cache replacement algorithm for the block cache
> --------------------------------------------------------------
>
> Key: KUDU-613
> URL: https://issues.apache.org/jira/browse/KUDU-613
> Project: Kudu
> Issue Type: Improvement
> Components: perf
> Affects Versions: M4.5
> Reporter: Andrew Wang
> Assignee: Mahesh Reddy
> Priority: Major
> Labels: performance, roadmap-candidate
>
> The block cache currently uses LRU, which is vulnerable to large scan
> workloads. It'd be good to implement something like 2Q.
> ARC (patent encumbered, but good for ideas):
> https://www.usenix.org/conference/fast-03/arc-self-tuning-low-overhead-replacement-cache
> HBase (2Q like):
> https://github.com/apache/hbase/blob/master/hbase-server/src/main/java/org/apache/hadoop/hbase/io/hfile/LruBlockCache.java
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
