Support EPoll/KQueue

[no2chem]

Overview

Description: This PR builds on top of #1061 to support Epoll/Kqueue server channel types.

Why should this be merged: This relatively simple addition on top of #1061 allows the use of native Netty transports, which should be more performant than the default NIO channel type. In addition, this change paves the way to use the built-in Local channel type, which will allow us to use Netty for all tests.

Checklist (Definition of Done):

• There are no TODOs left in the code
• Coding conventions (e.g. for logging, unit tests) have been followed
• Change is covered by automated tests
• Public API has Javadoc

[coveralls]

Changes Unknown when pulling c42bc87 on EPoll into ** on master**.

[codecov]

Codecov Report

❗ No coverage uploaded for pull request base (master@5b4ffd0). Click here to learn what that means.
The diff coverage is 31.57%.

@@            Coverage Diff            @@
##             master    #1064   +/-   ##
=========================================
  Coverage          ?   66.85%           
=========================================
  Files             ?      225           
  Lines             ?    10259           
  Branches          ?     1019           
=========================================
  Hits              ?     6859           
  Misses            ?     3007           
  Partials          ?      393

Impacted Files	Coverage Δ
...n/java/org/corfudb/infrastructure/CorfuServer.java	39.7% <31.57%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 5b4ffd0...5a55bab. Read the comment docs.

[corfudb-performance]

Results automatically generated by CorfuDB Benchmark Framework to assess the performance of this pull request for commit 1848e77.

*** 0.0% transaction FAILURE rate for NonConflictingTx+Scan workload, 1 threads, Disk mode
*** 0.0% transaction FAILURE rate for NonConflictingTx+Scan workload, 5 threads, Disk mode
*** 0.0% transaction FAILURE rate for NonConflictingTx+Scan workload, 10 threads, Disk mode
*** 0.0% transaction FAILURE rate for NonConflictingTx workload, 1 threads, Disk mode
*** 0.0% transaction FAILURE rate for NonConflictingTx workload, 5 threads, Disk mode
*** 0.0% transaction FAILURE rate for NonConflictingTx workload, 10 threads, Disk mode

An interactive dashboard with Pull Request Performance Metrics for ALL cluster types and numbers of threads in run, is available at:
Pull Request #1064 Graphs

[medhavidhawan]

I hope this is ok. It is a z change so should be fine but I am not sure.

[no2chem]

@medhavidhawan should we change the default to NIO? Or should we keep it auto?

[medhavidhawan]

I think we should keep the default to NIO. We can then test auto separately.

[corfudb-performance]

Results automatically generated by CorfuDB Benchmark Framework to assess the performance of this pull request for commit 328d59c.

*** 0.0% transaction FAILURE rate for NonConflictingTx+Scan workload, 1 threads, Disk mode
*** 0.0% transaction FAILURE rate for NonConflictingTx+Scan workload, 5 threads, Disk mode
*** 0.0% transaction FAILURE rate for NonConflictingTx+Scan workload, 10 threads, Disk mode
*** 0.0% transaction FAILURE rate for NonConflictingTx workload, 1 threads, Disk mode
*** 0.0% transaction FAILURE rate for NonConflictingTx workload, 5 threads, Disk mode
*** 0.0% transaction FAILURE rate for NonConflictingTx workload, 10 threads, Disk mode

An interactive dashboard with Pull Request Performance Metrics for ALL cluster types and numbers of threads in run, is available at:
Pull Request #1064 Graphs

[corfudb-performance]

Results automatically generated by CorfuDB Benchmark Framework to assess the performance of this pull request for commit 5a55bab.

*** 0.0% transaction FAILURE rate for NonConflictingTx+Scan workload, 1 threads, Disk mode
*** 0.0% transaction FAILURE rate for NonConflictingTx+Scan workload, 5 threads, Disk mode
*** 0.0% transaction FAILURE rate for NonConflictingTx+Scan workload, 10 threads, Disk mode
*** 0.0% transaction FAILURE rate for NonConflictingTx workload, 1 threads, Disk mode
*** 0.0% transaction FAILURE rate for NonConflictingTx workload, 5 threads, Disk mode
*** 0.0% transaction FAILURE rate for NonConflictingTx workload, 10 threads, Disk mode

An interactive dashboard with Pull Request Performance Metrics for ALL cluster types and numbers of threads in run, is available at:
Pull Request #1064 Graphs