Simplify bulk request execution #20109

areek · 2016-08-22T22:50:04Z

Currently, bulk item requests can be any ActionRequest, this PR
restricts bulk item requests to DocumentRequest. This simplifies
handling failures during bulk requests. Additionally, a new enum
is added to DocumentRequest to represent the intended operation
to be performed by a document request (create, index, update
and delete), which was previously represented with a mix of strings
and index request operation type.
Now, index request operation type reuses the new enum to specify
whether the request should create or index a document.
Restricting bulk requests to DocumentRequest further simplifies
execution of shard-level bulk operations to use the same failure
handling for index, delete and update operations.
This PR also fixes a bug which executed delete operations twice for
replica copies while executing bulk requests.

relates #19105

Currently, bulk item requests can be any ActionRequest, this commit restricts bulk item requests to DocumentRequest. This simplifies handling failures during bulk requests. Additionally, a new enum is added to DocumentRequest to represent the intended operation to be performed by a document request. Now, index operation type also uses the new enum to specify whether the request should create or index a document.

This commit refactors execution of shard-level bulk operations to use the same failure handling for index, delete and update operations.

bleskes

I went through this very carefully and it looks great. It's an amazing restructuring and I'm happy you took it on yourself. I left some minor comments. I also think we can potentially simplify further and make the DocumentRequest inherit from WriteReplicationRequest (and call DocWriteRequest) . I'm not sure but it would be great if you can give it a go.

bleskes · 2016-10-04T13:43:18Z

core/src/main/java/org/elasticsearch/action/bulk/BulkItemRequest.java

+            UpdateRequest updateRequest = new UpdateRequest();
+            updateRequest.readFrom(in);
+            request = updateRequest;
        }


add an else and throw an exception?

Thanks for the suggestion

bleskes · 2016-10-04T14:07:39Z

core/src/main/java/org/elasticsearch/action/DocumentRequest.java

+
+        OpType(int op) {
+            this.op = (byte) op;
+            this.lowercase = this.toString().toLowerCase(Locale.ENGLISH);


don't we typically use ROOT for these things?

changed it to ROOT, but there are a few places (e.g. DocWriteResponse) where we use ENGLISH too. maybe we should change it to ROOT as well?

bleskes · 2016-10-04T14:12:57Z

core/src/main/java/org/elasticsearch/action/bulk/BulkRequest.java

        out.writeVInt(requests.size());
-        for (ActionRequest<?> request : requests) {
+        for (DocumentRequest<?> request : requests) {
            if (request instanceof IndexRequest) {


shall we fold these reads and writes into static methods of DocumentRequest?

bleskes · 2016-10-04T14:15:13Z

core/src/main/java/org/elasticsearch/action/bulk/TransportBulkAction.java

+            if (request == null) {
+                continue;
            }
+            String concreteIndex = concreteIndices.getConcreteIndex(request.index()).getName();


bleskes · 2016-10-04T14:26:03Z

core/src/main/java/org/elasticsearch/action/bulk/TransportShardBulkAction.java

-                setResponse(item, item.getPrimaryResponse());
+            BulkItemRequest item = request.items()[requestIndex];
+            DocumentRequest<?> documentRequest = item.request();
+            if (ExceptionsHelper.status(e) == RestStatus.CONFLICT) {


can we stay consistent and use isConflictException? (I know it was like this before)

bleskes · 2016-10-05T12:52:17Z

core/src/main/java/org/elasticsearch/action/bulk/TransportShardBulkAction.java

-            location = locationToSync(location, result.getLocation());
+            WriteResult<? extends DocWriteResponse> writeResult = innerExecuteBulkItemRequest(metaData, indexShard,
+                request, requestIndex);
+            if (writeResult.getResponse().getResult() != DocWriteResponse.Result.NOOP) {


why not check if writeResult.getLocation() != null?

I wanted to be more strict as only a noop result can be valid with a null location. changed it to be an assert instead.

bleskes · 2016-10-05T13:15:41Z

core/src/main/java/org/elasticsearch/action/bulk/BulkItemResponse.java

    public void writeTo(StreamOutput out) throws IOException {
        out.writeVInt(id);
-        out.writeString(opType);
+        out.writeByte(opType.getId());


we need bwc here too

Currently, update action delegates to index and delete actions for replication using a dedicated transport action. This change makes update a replication operation, removing the dedicated transport action. This simplifies bulk execution and removes duplicate logic for update retries and translation. This consolidates the interface for single document write requests. Now on the primary, the update request is translated to an index or delete request before execution and the translated request is sent to copies for replication.

areek · 2016-10-06T09:20:42Z

Thanks @bleskes for the review :). I addressed all the minor comments.

I also think we can potentially simplify further and make the DocumentRequest inherit from WriteReplicationRequest (and call DocWriteRequest)

I really like the idea of making DocumentRequest inherit from WriteReplicationRequest. While giving this a go, I refactored update operation to be a replication operation (updates are a DocumentWriteRequest but does't use the replication operation instead delegates to index and delete operations for replication). This cleans up the dedicated transport update action (TransportInstanceSingleOperationAction) in favour of reusing replication action and duplicate code for update retry logic and translation. The one big downside is it would be difficult to implement wire bwc with 5.0. WDYT?

…ification" This reverts commit 2a651fc.

This reverts commit 42bc2d1.

This reverts commit b5079ce.

This reverts commit eee0d18.

bleskes

Thanks @areek . Let's leave the merger/inheritance of DocumentRequest and ReplicationRequest alone. None of the solutions seem ideal (update request is not a replication request and the fact it's running on the primary is not a good thing). This PR is great enough and it would be a shame to delay it. We can revisit / evaluate again when have a better idea.

I do have two small asks (next to a question I left in the comments)

Can we rename DocumentRequest to DocWriteRequest? it will be inline with the DocWriteResponse we have today.
It seems we can get rid of the <?> in all the DcoumentRequest<?> - I think it will make it simpler to read. See https://gist.github.com/bleskes/7ec26991d7be6c816103b8cac65353a4

bleskes · 2016-10-09T18:56:35Z

core/src/main/java/org/elasticsearch/action/bulk/TransportShardBulkAction.java

-                    return new UpdateResult(translate, indexRequest, retry, cause, null);
+                } else {
+                    assert translate.getResponseResult() == DocWriteResponse.Result.UPDATED;
+                    update.setGetResult(updateHelper.extractGetResult(updateRequest, updateRequest.concreteIndex(), indexResponse.getVersion(), translate.updatedSourceAsMap(), translate.updateSourceContentType(), indexSourceAsBytes));


why do we now return always the get result on updates? it seems to be different before

I changed it to match the logic in TransportUpdateAction.shardOperation but that caused CI failures, I reverted the change to be the same as update operation in bulk

bleskes · 2016-10-09T19:02:50Z

It also seems CI isn't happy.

areek · 2016-10-12T03:33:11Z

Thanks @bleskes for the feedback. I updated the PR, addressing your comments and the CI is happy. Could you take a look

bleskes · 2016-10-12T11:06:16Z

LGTM. Awesome @areek

This is a bespoke backport of elastic#20109 for 5.x: Currently, bulk item requests can be any ActionRequest, this PR restricts bulk item requests to DocumentRequest. This simplifies handling failures during bulk requests. Additionally, a new enum is added to DocumentRequest to represent the intended operation to be performed by a document request (create, index, update and delete), which was previously represented with a mix of strings and index request operation type. Now, index request operation type reuses the new enum to specify whether the request should create or index a document. Restricting bulk requests to DocumentRequest further simplifies execution of shard-level bulk operations to use the same failure handling for index, delete and update operations. This PR also fixes a bug which executed delete operations twice for replica copies while executing bulk requests. Relates to elastic#19105 and elastic#20109

areek added >enhancement review v5.0.0-beta1 labels Aug 22, 2016

areek force-pushed the cleanup/transport_bulk branch from ff103b3 to f207ecc Compare August 23, 2016 14:33

areek mentioned this pull request Sep 1, 2016

Simplify write failure handling #19105

Merged

areek added 2 commits September 1, 2016 14:15

Simplify shard-level bulk operation execution

cc993de

This commit refactors execution of shard-level bulk operations to use the same failure handling for index, delete and update operations.

Fix double delete on replica copy when executing bulk request

14908f8

areek force-pushed the cleanup/transport_bulk branch from f207ecc to 14908f8 Compare September 1, 2016 18:16

clintongormley added v5.0.0 and removed v5.0.0-beta1 labels Sep 14, 2016

areek force-pushed the cleanup/transport_bulk branch from 1fa265b to 5677cb6 Compare October 3, 2016 18:35

Merge branch 'master' into cleanup/transport_bulk

248ac24

areek force-pushed the cleanup/transport_bulk branch from 5677cb6 to 248ac24 Compare October 3, 2016 20:12

areek added 2 commits October 4, 2016 14:06

Merge branch 'master' into cleanup/transport_bulk

bd4a03a

ensure bwc wire compatibility

40b4f39

areek force-pushed the cleanup/transport_bulk branch from e515328 to 40b4f39 Compare October 4, 2016 19:05

bleskes suggested changes Oct 5, 2016

View reviewed changes

areek added 3 commits October 5, 2016 17:51

Merge branch 'master' into cleanup/transport_bulk

9b691f0

cleanup

57d8025

areek force-pushed the cleanup/transport_bulk branch from e68723a to eee0d18 Compare October 6, 2016 09:03

rename DocumentRequest to DocumentWriteRequest

b5079ce

areek force-pushed the cleanup/transport_bulk branch 2 times, most recently from 8f01c60 to 02ea4b8 Compare October 6, 2016 18:42

fix bug in bulk replication for noop update operation

42bc2d1

areek force-pushed the cleanup/transport_bulk branch from 02ea4b8 to 42bc2d1 Compare October 6, 2016 18:47

remove duplicate logic for request resolution and routing verification

2a651fc

areek added 7 commits October 7, 2016 17:48

Revert "remove duplicate logic for request resolution and routing ver…

5bbdcd6

…ification" This reverts commit 2a651fc.

Revert "fix bug in bulk replication for noop update operation"

68c82cd

This reverts commit 42bc2d1.

Revert "rename DocumentRequest to DocumentWriteRequest"

396f80c

This reverts commit b5079ce.

Revert "Make update a replication action"

97a6756

This reverts commit eee0d18.

Merge branch 'master' into cleanup/transport_bulk

c747085

remove redundant final qualifier

9d48248

fix bug in update operation in shard bulk execution

fe50db2

bleskes reviewed Oct 9, 2016

View reviewed changes

bleskes added v6.0.0-alpha1 and removed v5.0.0 labels Oct 9, 2016

areek added 4 commits October 11, 2016 14:21

fix update operation in bulk execution

225a04b

change DocumentRequest<?> to DocumentRequest for readibility

661067d

rename DocumentRequest to DocWriteRequest

0e8b653

Merge branch 'master' into cleanup/transport_bulk

481f790

bleskes approved these changes Oct 12, 2016

View reviewed changes

Merge branch 'master' into cleanup/transport_bulk

133be66

areek merged commit 133be66 into elastic:master Oct 12, 2016

imotov mentioned this pull request Dec 8, 2016

Add descriptions to bulk tasks #22059

Merged

dakrone mentioned this pull request Jan 19, 2017

Simplify bulk request execution #22697

Merged

lcawl added :Distributed Indexing/CRUD A catch all label for issues around indexing, updating and getting a doc by id. Not search. and removed :Bulk labels Feb 13, 2018

Simplify bulk request execution #20109

Simplify bulk request execution #20109

Uh oh!

Conversation

areek commented Aug 22, 2016

Uh oh!

bleskes left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

areek commented Oct 6, 2016

Uh oh!

bleskes left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

bleskes commented Oct 9, 2016

Uh oh!

areek commented Oct 12, 2016

Uh oh!

bleskes commented Oct 12, 2016

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants