Azure batch deletes

[lightningrob]

No description provided.

[ankagrawal]

can the statuscode returned by exception be one of OK, ACCEPTED, NOT_FOUND or GONE? If not, then maybe we can move the try catch block to include switch as well.

[lightningrob]

They throw exception if the http status is not the "expected" value, which in this case is Accepted.

[codecov-io]

Codecov Report

❗ No coverage uploaded for pull request base (master@805a993). Click here to learn what that means.
The diff coverage is 91.17%.

@@            Coverage Diff            @@
##             master    #1365   +/-   ##
=========================================
  Coverage          ?   71.99%           
  Complexity        ?     6759           
=========================================
  Files             ?      486           
  Lines             ?    38378           
  Branches          ?     4880           
=========================================
  Hits              ?    27629           
  Misses            ?     9415           
  Partials          ?     1334

Impacted Files	Coverage Δ	Complexity Δ
...b.ambry.clustermap/StaticClusterAgentsFactory.java	31.57% <0%> (ø)	4 <0> (?)
...ithub.ambry.cloud/azure/AzureCloudDestination.java	72.98% <100%> (ø)	30 <0> (?)
.../main/java/com.github.ambry/store/MessageInfo.java	89.41% <100%> (ø)	25 <1> (?)
...ls/src/main/java/com.github.ambry.utils/Utils.java	78.46% <100%> (ø)	124 <5> (?)
...essageformat/UndeleteMessageFormatInputStream.java	100% <100%> (ø)	1 <1> (?)
...b.ambry.clustermap/SimpleClusterChangeHandler.java	91.78% <100%> (ø)	50 <0> (?)
...java/com.github.ambry/config/ClusterMapConfig.java	100% <100%> (ø)	3 <0> (?)
...hub.ambry/clustermap/StateTransitionException.java	100% <100%> (ø)	2 <0> (?)
...a/com.github.ambry.clustermap/ClusterMapUtils.java	93.86% <100%> (ø)	46 <4> (?)
...java/com.github.ambry.router/GetBlobOperation.java	91.52% <100%> (ø)	44 <0> (?)
... and 7 more

---

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 805a993...06e5b3f. Read the comment docs.

[cgtz]

LGTM after comments are addressed

[cgtz]

could this loop be simplified to:

for (int start = 0; start < inputList.size(); start += batchSize) {
  int end = Math.min(start + batchSize, inputList.size());
  partitionedLists.add(inputList.subList(start, end));
}

I may be missing an edge case here though

[cgtz]

So I was thinking about this partition thing and it seems that we don't really need a full list and just need an iterable that we can do a for loop on. Something like this could work, I think:

  public static <T> Iterable<List<T>> partition(List<T> list, int batchSize) {
    return () -> new Iterator<List<T>>() {
      int nextStart = 0;

      @Override
      public boolean hasNext() {
        return nextStart < list.size();
      }

      @Override
      public List<T> next() {
        if (!hasNext()) {
          throw new NoSuchElementException();
        }
        int start = nextStart;
        nextStart += batchSize;
        return list.subList(start, Math.min(start + batchSize, list.size()));
      }
    };
  }

[lightningrob]

I'm not sure this is a big improvement. I prefer to mimic the guava utility. Once we upgrade to a newer JDK, we can probably retire the utility and use the Collections one.

[lightningrob]

I made the simplification you suggested.