[
https://issues.apache.org/jira/browse/HDFS-17041?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17730258#comment-17730258
]
ASF GitHub Bot commented on HDFS-17041:
---------------------------------------
virajjasani commented on code in PR #5723:
URL: https://github.com/apache/hadoop/pull/5723#discussion_r1222057296
##########
hadoop-hdfs-project/hadoop-hdfs-rbf/src/main/java/org/apache/hadoop/hdfs/server/federation/store/driver/impl/StateStoreFileBaseImpl.java:
##########
@@ -400,10 +403,8 @@ public <T extends BaseRecord> StateStoreOperationResult
putAll(
toWrite.put(recordPath, record);
} else if (errorIfExists) {
LOG.error("Attempt to insert record {} that already exists",
recordPath);
- if (metrics != null) {
- metrics.addFailure(monotonicNow() - start);
Review Comment:
We will have metrics, at the end
> RBF: Fix putAll impl for mysql and file based state stores
> ----------------------------------------------------------
>
> Key: HDFS-17041
> URL: https://issues.apache.org/jira/browse/HDFS-17041
> Project: Hadoop HDFS
> Issue Type: Bug
> Reporter: Viraj Jasani
> Assignee: Viraj Jasani
> Priority: Major
> Labels: pull-request-available
>
> Only zookeeper based state store allows all records to be inserted even
> though only few of them already exists and "errorIfExists" is true, however
> file/fs as well as mysql based putAll fails the whole putAll operation
> immediately after encountering single record that already exists in the
> records and "errorIfExists" is provided true (which is the case while
> inserting records for the first time).
> For all implementations, we should allow inserts of the records that do not
> already exist and report any record as failure that already exists, rather
> than failing the whole operation and not trying to insert valid records.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
