Fix broken backward compatibility from 2.7 for IndexSorted field indices

[gashutos]

Description

Upgrade to OpenSearch 2.7 is breaking with change #6424 . Not for all scenarios but when any index contains IndexSorted field of either BYTE, BOOLEAN, SHORT, INTEGER, where the IndexSortField we created in prior version was upcasted to LONG for all above types, which was removed in #6424 to achieve better performance for sort queries.
The SortField for IndexSort is stored in SegmentInfo and that gets retained when index is migrating from old cluster to new 2.7 cluster, and hence in IndexWriter creation, Lucene throws Exception because on 2.7, IndexWriteConfiguration has IndexSort with non-upcasted type while in SegmentInfo Lucene finds the same sort field with up-casted.

i.e, check below scenario, where Index sort is on Integer field, and the index is migrating from older version to 2.7 version, and on newer version, the IndexWriterCreation fails with IndexSort mismatch between IndexWriteIndexSort & Lucene segmentinfo contained indexsort.

"sort" : {
          "field" : "position",
          "order" : "desc"
        },
"position" : {
          "type" : "integer"
        },


Caused by: RemoteTransportException[[d3aea48e6d44f0fc824a9051f26d9969][:9300][internal:index/shard/recovery/prepare_translog]]; nested: IllegalArgumentException[cannot change previous indexSort=<sortednumeric: "position">! missingValue=-9223372036854775808 selector=MAX type=LONG (from segment=_8g9x(9.4.2):C61519:[indexSort=<sortednumeric: "position">! missingValue=-9223372036854775808 selector=MAX type=LONG]:[diagnostics={os=Linux, os.version=5.10.149-133.644.amzn2.aarch64, timestamp=1694560481039, mergeMaxNumSegments=-1, lucene.version=9.4.2, source=merge, os.arch=aarch64, java.runtime.version=11.0.18+10-LTS, mergeFactor=4, java.vendor=Amazon.com Inc.}]:[attributes={Lucene90StoredFieldsFormat.mode=BEST_SPEED}]:fieldInfosGen=36:dvGen=36 :softDel=27709 :id=5gdgx085ys6f3xyyb68rpx85c) to new indexSort=<sortednumeric: "position">! missingValue=-2147483648 selector=MAX type=INT];
Caused by: java.lang.IllegalArgumentException: cannot change previous indexSort=<sortednumeric: "position">! missingValue=-9223372036854775808 selector=MAX type=LONG (from segment=_8g9x(9.4.2):C61519:[indexSort=<sortednumeric: "position">! missingValue=-9223372036854775808 selector=MAX type=LONG]:[diagnostics={os=Linux, os.version=5.10.149-133.644.amzn2.aarch64, timestamp=1694560481039, mergeMaxNumSegments=-1, lucene.version=9.4.2, source=merge, os.arch=aarch64, java.runtime.version=11.0.18+10-LTS, mergeFactor=4, java.vendor=Amazon.com Inc.}]:[attributes={Lucene90StoredFieldsFormat.mode=BEST_SPEED}]:fieldInfosGen=36:dvGen=36 :softDel=27709 :id=5gdgx085ys6f3xyyb68rpx85c) to new indexSort=<sortednumeric: "position">! missingValue=-2147483648 selector=MAX type=INT
        at org.apache.lucene.index.IndexWriter.validateIndexSort(IndexWriter.java:1208)
        at org.apache.lucene.index.IndexWriter.<init>(IndexWriter.java:1119)
        at org.opensearch.index.engine.InternalEngine.createWriter(InternalEngine.java:2437)
        at org.opensearch.index.engine.InternalEngine.createWriter(InternalEngine.java:2425)
        at org.opensearch.index.engine.InternalEngine.<init>(InternalEngine.java:305)
        at org.opensearch.index.engine.InternalEngine.<init>(InternalEngine.java:223)
        at org.opensearch.index.engine.InternalEngineFactory.newReadWriteEngine(InternalEngineFactory.java:43)
        at org.opensearch.index.shard.IndexShard.innerOpenEngineAndTranslog(IndexShard.java:2286)
        at org.opensearch.index.shard.IndexShard.openEngineAndSkipTranslogRecovery(IndexShard.java:2259)
        at org.opensearch.indices.recovery.RecoveryTarget.lambda$prepareForTranslogOperations$1(RecoveryTarget.java:256)

I have added check to treat old migrating indices to treat as Sort field type upcatsed when they are moving to OpenSearch 2.7.

Related Issues

Resolves #10044

Testing done

• Tested below scenarions for all BYTE, BOOLEAN, INTEGER, SHORT types with Index sort
• Migrating indices from OpenSearch 2.5 to OpenSearch 2.7
• Indexing more documents to above migrated indices in OpenSearch 2.7
• Fresh index on OpenSearch 2.7
• Old index created on OpenSearch 2.7 iteself and upgraded to new OpenSearch 2.7 version with this patch
• Sort queries on all above scenarios in asc/desc order

Check List

• New functionality includes testing.
  • All tests pass
• New functionality has been documented.
  • New functionality has javadoc added
• Commits are signed per the DCO using --signoff
• Commit changes are listed out in CHANGELOG.md file (See: Changelog)

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.
For more information on following Developer Certificate of Origin and signing off your commits, please check here.

[github-actions]

Gradle Check (Jenkins) Run Completed with:

• RESULT: SUCCESS ✅
• URL: https://build.ci.opensearch.org/job/gradle-check/25681/
• CommitID: f61aba6

[reta]

@Bukhtawar all cleared out from your side? thanks!

[Bukhtawar]

Do we need to have a BWC IT test for this?

[gashutos]

Do we need to have a BWC IT test for this?

I have added this to mixed-cluster, where Index sorted indices will be having replicas on multiple versions, that should suffice this scenario.

[github-actions]

Gradle Check (Jenkins) Run Completed with:

• RESULT: FAILURE ❌
• URL: https://build.ci.opensearch.org/job/gradle-check/25691/
• CommitID: 9393aba
  Please examine the workflow log, locate, and copy-paste the failure(s) below, then iterate to green.
  Is the failure a flaky test unrelated to your change?

[github-actions]

Gradle Check (Jenkins) Run Completed with:

• RESULT: UNSTABLE ❕
• TEST FAILURES:

      1 org.opensearch.search.SearchWeightedRoutingIT.testMultiGetWithNetworkDisruption_FailOpenEnabled

• URL: https://build.ci.opensearch.org/job/gradle-check/25694/
• CommitID: 61d355e
  Please review all flaky tests that succeeded after retry and create an issue if one does not already exist to track the flaky failure.

[gashutos]

Thanks for reviews @reta @backslasht @Bukhtawar

@reta help me merging this once @backslasht approves.

[backslasht]

Looks good to me! Thanks @gashutos for quickly fixing it.

[github-actions]

Gradle Check (Jenkins) Run Completed with:

• RESULT: SUCCESS ✅
• URL: https://build.ci.opensearch.org/job/gradle-check/25701/
• CommitID: af7e9e9

[opensearch-trigger-bot]

The backport to 2.10 failed:

The process '/usr/bin/git' failed with exit code 128

To backport manually, run these commands in your terminal:

# Navigate to the root of your repository
cd $(git rev-parse --show-toplevel)
# Fetch latest updates from GitHub
git fetch
# Create a new working tree
git worktree add ../.worktrees/OpenSearch/backport-2.10 2.10
# Navigate to the new working tree
pushd ../.worktrees/OpenSearch/backport-2.10
# Create a new branch
git switch --create backport/backport-10045-to-2.10
# Cherry-pick the merged commit of this pull request and resolve the conflicts
git cherry-pick -x --mainline 1 d34b35272fa19bc53cb0f6843a92d2d1ee4e3b47
# Push it to GitHub
git push --set-upstream origin backport/backport-10045-to-2.10
# Go back to the original working tree
popd
# Delete the working tree
git worktree remove ../.worktrees/OpenSearch/backport-2.10

Then, create a pull request where the base branch is 2.10 and the compare/head branch is backport/backport-10045-to-2.10.