[OOZIE-1893] Recovery service will never recover bundle action if CoordSubmitXCommand command is lost

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 4.1.0
• Assignee: Purshotam Shah
• Reporter: Purshotam Shah
• Created: 19/Jun/14 22:45
• Updated: 09/Dec/14 00:45
• Resolved: 23/Jun/14 21:21

Description

If bundle action is in prep and coord submit command is lost, bundle action will never recover bcz of below condition.

baction.getCoordId() is null in that case and isJobIdForThisServer(baction.getCoordId()) will throw NPE.

RecoveryService.runBundleRecovery()

if (baction.getCoordId() == null && baction.getStatus() != Job.Status.PREP) {
	log.error("CoordId is null for Bundle action " + baction.getBundleActionId());
	continue;
}
if (Services.get().get(JobsConcurrencyService.class).isJobIdForThisServer(baction.getCoordId())) {
}