Persistance concurrency bug

[yojs]

Issue #: #324

Description of changes:
This change makes us print less stacktrace on errors that are expected. Also resolves a concurrency bug while threads doing writes and while the file is being rotated.

Tests:
Added a test

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.

[codecov]

Codecov Report

Merging #323 into master will decrease coverage by 0.27%.
The diff coverage is 44.59%.

@@             Coverage Diff              @@
##             master     #323      +/-   ##
============================================
- Coverage     67.15%   66.88%   -0.28%     
+ Complexity     2034     2033       -1     
============================================
  Files           297      297              
  Lines         13063    13095      +32     
  Branches       1081     1086       +5     
============================================
- Hits           8773     8759      -14     
- Misses         3910     3954      +44     
- Partials        380      382       +2     

Impacted Files	Coverage Δ	Complexity Δ
...erformanceanalyzer/rca/persistence/FileRotate.java	52.94% <16.66%> (-17.06%)	10.00 <1.00> (+1.00)	⬇️
...manceanalyzer/rca/persistence/SQLitePersistor.java	72.98% <52.38%> (-7.77%)	23.00 <0.00> (ø)
...ormanceanalyzer/rca/persistence/PersistorBase.java	82.60% <85.71%> (-6.87%)	20.00 <0.00> (ø)
...ch/performanceanalyzer/rca/persistence/FileGC.java	92.72% <100.00%> (ø)	14.00 <0.00> (ø)
...rch/performanceanalyzer/rca/framework/api/Rca.java	80.00% <0.00%> (-14.29%)	5.00% <0.00%> (ø%)
.../performanceanalyzer/rca/framework/core/Stats.java	97.82% <0.00%> (-2.18%)	22.00% <0.00%> (-1.00%)
...csearch/performanceanalyzer/rca/RcaController.java	80.76% <0.00%> (-0.55%)	38.00% <0.00%> (-1.00%)

---

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 1ac07f1...ed3410c. Read the comment docs.

[rguo-aws]

I assume creating the SQL statement itself does not throw any exception. Can we leave this part out of the try block ?

[yojs]

The DSLContext can be uninitialized if we are not using synchronization between the threads. I saw a case where the table was added to hastable but the columns for the table were missing. We gather some date to create the createStatement and it the gathering part we can run into exceptions.
These cases are fixed now but it would be good to catch and log exception. But you are right, the scope of exceptions is larger and so I have added code to catch exception and not just DataAcessException.