[
https://issues.apache.org/jira/browse/GOBBLIN-1875?focusedWorklogId=876235&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-876235
]
ASF GitHub Bot logged work on GOBBLIN-1875:
-------------------------------------------
Author: ASF GitHub Bot
Created on: 15/Aug/23 00:24
Start Date: 15/Aug/23 00:24
Worklog Time Spent: 10m
Work Description: umustafi commented on code in PR #3737:
URL: https://github.com/apache/gobblin/pull/3737#discussion_r1294080873
##########
gobblin-runtime/src/main/java/org/apache/gobblin/scheduler/JobScheduler.java:
##########
@@ -586,9 +587,13 @@ public void close() throws IOException {
* Get a {@link org.quartz.Trigger} from the given job configuration
properties.
*/
public static Trigger createTriggerForJob(JobKey jobKey, Properties
jobProps) {
- // Build a trigger for the job with the given cron-style schedule
+ Random random = new Random();
+ /*
+ Build a trigger for the job with the given cron-style schedule
+ Append a random integer to job name to be able to add multiple triggers
associated with the same job.
+ */
return TriggerBuilder.newTrigger()
- .withIdentity(jobProps.getProperty(ConfigurationKeys.JOB_NAME_KEY),
+ .withIdentity(jobProps.getProperty(ConfigurationKeys.JOB_NAME_KEY) +
random.nextInt(1000),
Review Comment:
updated with this!
Issue Time Tracking
-------------------
Worklog Id: (was: 876235)
Time Spent: 50m (was: 40m)
> Create Unique Trigger identifiers
> ----------------------------------
>
> Key: GOBBLIN-1875
> URL: https://issues.apache.org/jira/browse/GOBBLIN-1875
> Project: Apache Gobblin
> Issue Type: Bug
> Components: gobblin-service
> Reporter: Urmi Mustafi
> Assignee: Abhishek Tiwari
> Priority: Major
> Time Spent: 50m
> Remaining Estimate: 0h
>
> We are seeing the following errors when trying to schedule reminders for
> Multi-active scheduler after losing lease acquisition:
> {{Unable to store Trigger with name: '...' and group: '...' because one
> already exists with this identification}}
> The Quartz scheduler allows multiple unique triggers mapping to one job in a
> N:1 relation, but this requires the trigger having a unique identifier
> associated with it. When we create triggers, we were previously using only
> the job name and group to identify the trigger so we're unable to add
> additional triggers for the same job. The trigger key is only used internally
> by the Quartz scheduler, so it is safe to manipulate by adding a random
> number to identify the trigger. We also don't expect more than a handful of
> triggers associated with a job, so an int 0 to 1000 will be sufficient to
> uniquely identify the trigger.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
