[
https://issues.apache.org/jira/browse/GOBBLIN-2049?focusedWorklogId=915410&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-915410
]
ASF GitHub Bot logged work on GOBBLIN-2049:
-------------------------------------------
Author: ASF GitHub Bot
Created on: 18/Apr/24 21:53
Start Date: 18/Apr/24 21:53
Worklog Time Spent: 10m
Work Description: Will-Lo commented on code in PR #3929:
URL: https://github.com/apache/gobblin/pull/3929#discussion_r1571428090
##########
gobblin-data-management/src/main/java/org/apache/gobblin/data/management/copy/writer/FileAwareInputStreamDataWriter.java:
##########
@@ -353,7 +357,8 @@ public static Path getOutputDir(State state) {
* Sets the {@link FsPermission}, owner, group for the path passed. It will
not throw exceptions, if operations
* cannot be executed, will warn and continue.
*/
- public static void safeSetPathPermission(FileSystem fs, FileStatus file,
OwnerAndPermission ownerAndPermission) {
+ public static void safeSetPathPermission(FileSystem fs, FileStatus file,
OwnerAndPermission ownerAndPermission,
Review Comment:
We should either rename this function, or create a new function, also if
it's the former also change the java doc. Currently the API documentation
states that it will explicitly warn and continue if operations cannot be
executed (which makes sense in some scenarios where permissions should not
cause the entire operation to fail), hence `safe` in the function name
Issue Time Tracking
-------------------
Worklog Id: (was: 915410)
Time Spent: 20m (was: 10m)
> Configure Gobblin Distcp Writer to fail if setPermission fails
> --------------------------------------------------------------
>
> Key: GOBBLIN-2049
> URL: https://issues.apache.org/jira/browse/GOBBLIN-2049
> Project: Apache Gobblin
> Issue Type: New Feature
> Components: gobblin-service
> Reporter: Urmi Mustafi
> Assignee: Abhishek Tiwari
> Priority: Major
> Time Spent: 20m
> Remaining Estimate: 0h
>
> Gobblin {{safeSetPathPermission}} does not throw an exception when setting
> permissions fail on a path. We want to change this behavior especially for
> use cases like manifest distcp where hundreds of thousands of files are
> involved in a distcp job and permission settings are important to replicate
> correctly as they cannot be updated or verified by hand.
> This PR adds a new configuration to fail writing or publishing tasks when the
> job is configured to report success when permissions are not replicated
> properly. The default behavior remains the same as before to allow the job to
> succeed without this for backwards compatibility.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
