[
https://issues.apache.org/jira/browse/HBASE-20188?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16418413#comment-16418413
]
stack commented on HBASE-20188:
-------------------------------
>From the report
>https://docs.google.com/spreadsheets/d/1w2NBqAPFthG8Ib4C0pHpLARYpWoIF2Vck2vHZW77zE4/edit#gid=0
>
{quote}Objective: Test base YCSB runs -- load, workloada, and workload b --
comparing hbase1.x to hbase2.x with and without IN_MEMORY_COMPACTION.
Description: Run three ycsb workloads: load, workloada (50/50), and workloadc
(100% reads) back to back. Zipfian. Client on one server, all regions on a
single RegionServer. Each workload runs for 20mins. Configs are default.There
are five columns to compare on: 1.2.7 vs. 2.0.0 vs 2.0.0 with no in-memory
compaction vs 2.0.0 with MSLAB disabled vs 2.0.0 with no in-memory compaction
and MSLAB disabled. Graphs on end of the page show four of the five runs.
Findings: As of now, hbase1.x performs better than 2.x in pure read and pure
write modes. Mixed load (workloada 50/50), 2.x is better. If
IN_MEMORY_COMPACTION is off, it is better still. IN_MEMORY_COMPACTION helps
pure writes slightly in 2.x but is detrimental in mixed load and when pure
read. {quote}
> [TESTING] Performance
> ---------------------
>
> Key: HBASE-20188
> URL: https://issues.apache.org/jira/browse/HBASE-20188
> Project: HBase
> Issue Type: Umbrella
> Components: Performance
> Reporter: stack
> Priority: Blocker
> Fix For: 2.0.0
>
> Attachments: ITBLL2.5B_1.2.7vs2.0.0_cpu.png,
> ITBLL2.5B_1.2.7vs2.0.0_gctime.png, ITBLL2.5B_1.2.7vs2.0.0_iops.png,
> ITBLL2.5B_1.2.7vs2.0.0_load.png, ITBLL2.5B_1.2.7vs2.0.0_memheap.png,
> ITBLL2.5B_1.2.7vs2.0.0_memstore.png, ITBLL2.5B_1.2.7vs2.0.0_ops.png,
> ITBLL2.5B_1.2.7vs2.0.0_ops_NOT_summing_regions.png, YCSB_CPU.png,
> YCSB_GC_TIME.png, YCSB_IN_MEMORY_COMPACTION=NONE.ops.png, YCSB_MEMSTORE.png,
> YCSB_OPs.png, YCSB_in-memory-compaction=NONE.ops.png, YCSB_load.png,
> flamegraph-1072.1.svg, flamegraph-1072.2.svg, tree.txt
>
>
> How does 2.0.0 compare to old versions? Is it faster, slower? There is rumor
> that it is much slower, that the problem is the asyncwal writing. Does
> in-memory compaction slow us down or speed us up? What happens when you
> enable offheaping?
> Keep notes here in this umbrella issue. Need to be able to say something
> about perf when 2.0.0 ships.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
