[SLING-4765] Add debugging info to duplicate sling.id detection in discovery.impl

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: Discovery Impl 1.1.2
• Fix Version/s: Discovery Impl 1.1.4
• Component/s: Extensions
• Assignee: Stefan Egli
• Reporter: Stefan Egli
• Created: 29/May/15 14:17
• Updated: 10/Jun/15 13:14
• Resolved: 04/Jun/15 11:11

Description

In case of SLING-2901 or SLING-2892 it sometimes is not exactly clear if indeed a second instance was hooked to the same repository or whether there is something else causing the detection to report a false positive.

To clarify this, add more debugging infos to the log output when these two situations occur, which should include:

• log runtimeId at activate time
• besides runtimeId, add pid/ip/port information which can then be logged in case a duplicate situation is detected (to simply proof which other 'ghost' instance was causing the duplicate situation)
• when the duplicate situation is detected, print all available information: local runtimeId/lastHeartbeat/pid/ip/port plus the same for what is stored under /var/discovery/impl/clusterInstances/<slingId>

Basically provide enough infos to proof the ghost duplicate instance (or help detecting that the issue is locally if it is)

Issue Links

• is related to: SLING-2892 Add simple paranoia check to detect duplicate sling.id in a cluster
• is related to: SLING-2901 Add robust paranoia check to detect duplicate sling.id in a cluster