[ISSUE-339] Optimize retry logic in send shuffle data

[xianjingfeng]

What changes were proposed in this pull request?

1. Set the default value of rss.client.retry.max to 50
2. Set rss.client.retry.max to rss.client.retry.max/2 if replica > replicaWrite
3. Throw an exception if rss.client.retry.max * rss.client.retry.interval.max > rss.client.send.check.timeout.ms

Why are the changes needed?

More reasonable. #339

Does this PR introduce any user-facing change?

No

How was this patch tested?

Added

[codecov-commenter]

Codecov Report

Merging #361 (62e7ff7) into master (57a834b) will increase coverage by 0.13%.
The diff coverage is 86.95%.

@@             Coverage Diff              @@
##             master     #361      +/-   ##
============================================
+ Coverage     58.46%   58.59%   +0.13%     
- Complexity     1572     1579       +7     
============================================
  Files           193      193              
  Lines         10846    10869      +23     
  Branches        954      956       +2     
============================================
+ Hits           6341     6369      +28     
+ Misses         4131     4126       -5     
  Partials        374      374              

Impacted Files	Coverage Δ
...apache/hadoop/mapreduce/v2/app/RssMRAppMaster.java	0.00% <0.00%> (ø)
...e/uniffle/client/factory/ShuffleClientFactory.java	0.00% <0.00%> (ø)
...rg/apache/uniffle/client/util/RssClientConfig.java	0.00% <ø> (ø)
...n/java/org/apache/hadoop/mapreduce/RssMRUtils.java	50.45% <100.00%> (+6.45%)	⬆️
...org/apache/spark/shuffle/RssSparkShuffleUtils.java	68.42% <100.00%> (+11.27%)	⬆️

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[jerqi]

LGTM, thanks @xianjingfeng , wait for CI

[jerqi]

Merged.