[
https://issues.apache.org/jira/browse/GOBBLIN-2050?focusedWorklogId=915641&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-915641
]
ASF GitHub Bot logged work on GOBBLIN-2050:
-------------------------------------------
Author: ASF GitHub Bot
Created on: 19/Apr/24 23:23
Start Date: 19/Apr/24 23:23
Worklog Time Spent: 10m
Work Description: phet commented on code in PR #3931:
URL: https://github.com/apache/gobblin/pull/3931#discussion_r1573050236
##########
gobblin-cluster/src/main/java/org/apache/gobblin/cluster/GobblinClusterConfigurationKeys.java:
##########
@@ -50,6 +50,8 @@ public class GobblinClusterConfigurationKeys {
public static final boolean DEFAULT_STANDALONE_CLUSTER_MODE = false;
// Root working directory for Gobblin cluster
public static final String CLUSTER_WORK_DIR = GOBBLIN_CLUSTER_PREFIX +
"workDir";
+ // Root working dir without appending the application name, keeping
CLUSTER_WORK_DIR property for backward compatibility
+ public static final String CLUSTER_ABSOLUTE_WORK_DIR =
GOBBLIN_CLUSTER_PREFIX + "absolute.workDir";
Review Comment:
not sure if I've missed some critical details here... but who's appending
the job_name to this prop?
in general, my pref wouldn't be to push the uniqueness requirement to the
caller, but rather to encapsulate within here the determination of a job-unique
dir, by taking in a prop for the "base dir" and then auto-appending the
particular job name to that
##########
gobblin-cluster/src/main/java/org/apache/gobblin/cluster/GobblinClusterConfigurationKeys.java:
##########
@@ -50,6 +50,8 @@ public class GobblinClusterConfigurationKeys {
public static final boolean DEFAULT_STANDALONE_CLUSTER_MODE = false;
// Root working directory for Gobblin cluster
public static final String CLUSTER_WORK_DIR = GOBBLIN_CLUSTER_PREFIX +
"workDir";
+ // Root working dir without appending the application name, keeping
CLUSTER_WORK_DIR property for backward compatibility
+ public static final String CLUSTER_ABSOLUTE_WORK_DIR =
GOBBLIN_CLUSTER_PREFIX + "absolute.workDir";
Review Comment:
not sure if I've missed some critical details here... but who's appending
the job_name to this prop?
in general, my pref wouldn't be to push the uniqueness requirement to the
caller, but rather to encapsulate within here the determination of a job-unique
dir, by first taking in a prop for the "base dir" and then auto-appending the
particular job name to that
Issue Time Tracking
-------------------
Worklog Id: (was: 915641)
Time Spent: 1h 40m (was: 1.5h)
> Allow configurable token file paths and cluster work directories to allow for
> easy cleanup after yarn app closes
> ----------------------------------------------------------------------------------------------------------------
>
> Key: GOBBLIN-2050
> URL: https://issues.apache.org/jira/browse/GOBBLIN-2050
> Project: Apache Gobblin
> Issue Type: Improvement
> Reporter: William Lo
> Priority: Major
> Time Spent: 1h 40m
> Remaining Estimate: 0h
>
> Gobblin Yarn Application Launcher has some issues where directories used for
> the job can persist after the job ends. It also creates a number of temporary
> files which can grow out of control.
> We want to be able to:
> 1. Clean up directories effectively
> 2. Use explicit paths to allow for consolidation of temp files to be under
> the same folder for token renewal and cluster work files.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
