[KUDU-1565] ksck should include some tserver health checks

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Component/s: ksck
• Labels: supportability
• Assignee: Unassigned
• Reporter: Todd Lipcon
• Created: 17/Aug/16 05:51
• Updated: 01/Jun/20 18:40

Description

KUDU-1516 added a few features to ksck, but still more improvements should be done to make this tool really solid for troubleshooting:

• as mentioned in the original issue, should check for raft configurations where at least one of the servers think there isn't a successful leader
• should check for replicas with transactions that have been running longer than some reasonable threshold (say 30 seconds)