[
https://issues.apache.org/jira/browse/OAK-6912?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16243648#comment-16243648
]
Andrei Dulceanu commented on OAK-6912:
--------------------------------------
[~frm], here's the output obtained when running the benchmark on both branches
(Oak 1.6 r1793526 [0] and Oak 1.6 r1793527 [1]) with your change for
{{SegmentCache}}:
|Oak 1.6 r1793527 (20170502)|{noformat}
# BasicWriteTest C min 10% 50% 90% max
N
Oak-Segment-Tar-Cold 1 59 62 66 76 918
824
{noformat}
{noformat}
CacheStats{hitCount=4564535, hitRate=1.00, missCount=0, missRate=0.00,
requestCount=4564535, loadCount=0, loadSuccessCount=0, loadExceptionCount=0,
totalLoadTime=0 min, 0 sec, averageLoadPenalty=0.00 ns, evictionCount=0,
elementCount=26, totalWeight=4.3 MB, maxWeight=104.9 MB}
{noformat}|
|Oak 1.6 r1793526 (20170502)|{noformat}
# BasicWriteTest C min 10% 50% 90% max
N
Oak-Segment-Tar-Cold 1 59 62 65 73 272
868
{noformat}
{noformat}
CacheStats{hitCount=4802390, hitRate=1.00, missCount=0, missRate=0.00,
requestCount=4802390, loadCount=0, loadSuccessCount=0, loadExceptionCount=0,
totalLoadTime=0 min, 0 sec, averageLoadPenalty=0.00 ns, evictionCount=0,
elementCount=27, totalWeight=4.5 MB, maxWeight=104.9 MB}
{noformat}|
The only difference between the two branches is the sequence of operations in
{{SegmentCache#put}}. While r1793526, first calls {{SegmentId#loaded}} and then
puts the segment in the cache, r1793527 does the other way round. Another thing
that influences the numbers is the fact that the new implementation for
{{SegmentCache#getSegment}} proposed by [~frm] loads first the segment and then
calls {{SegmentId#loaded}}, which might explain the almost identical cache
stats and durations recorded.
[0]
https://github.com/dulceanu/jackrabbit-oak/tree/issues/1_6/20170502/r1793526_frm
[1]
https://github.com/dulceanu/jackrabbit-oak/tree/issues/1_6/20170502/r1793527_frm
/cc [~mduerig]
> Cold standby performance regression due to segment caching
> ----------------------------------------------------------
>
> Key: OAK-6912
> URL: https://issues.apache.org/jira/browse/OAK-6912
> Project: Jackrabbit Oak
> Issue Type: Bug
> Components: segment-tar, tarmk-standby
> Affects Versions: 1.7.0
> Reporter: Andrei Dulceanu
> Assignee: Andrei Dulceanu
> Labels: cold-standby, performance, scalability
> Fix For: 1.8, 1.7.12
>
>
> The changes to the segment cache introduced in r1793527 [0] introduced a
> performance regression on the primary for the case in which a standby is
> attached to it. Below a benchmark duration comparison between primary w/o and
> w/ standby for r1793527 (after the segment cache changes) and r1793526
> (before the changes) :
> |Oak 1.6 r1793527 (20170502)|{noformat}
> # BasicWriteTest C min 10% 50% 90% max
> N
> Oak-Segment-Tar 1 19 21 22 26 160
> 2491
> Oak-Segment-Tar-DS 1 56 59 63 70 181
> 919
> Oak-Segment-Tar-Cold(Shared DS) 1 58 66 159 177 372
> 302
> {noformat}|
> |Oak 1.6 r1793526 (20170502)|{noformat}
> # BasicWriteTest C min 10% 50% 90% max
> N
> Oak-Segment-Tar 1 19 21 22 25 52
> 2584
> Oak-Segment-Tar-DS 1 56 60 63 69 158
> 925
> Oak-Segment-Tar-Cold(Shared DS) 1 57 60 64 70 122
> 915
> {noformat}|
> [0]
> https://github.com/apache/jackrabbit-oak/commit/efafa4e1710621b7f3b8e92d0b2681669185fcd4
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
