[OAK-2108] Killing a node may stop async index update to to 30 minutes (Tar storage)

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.5.14, 1.6.0
• Component/s: query
• Labels: resilience
• Epic Link: indexer resilience 1.6
• Assignee: Chetan Mehrotra
• Reporter: Thomas Mueller
• Created: 18/Sep/14 09:27
• Updated: 26/Sep/19 11:39
• Resolved: 20/Nov/16 16:31

Description

When killing a node that is running the sync index update, then this async index update will not run for up to 15 minutes, because the lease time is set to 15 minutes.

I think the lease time should be much smaller, for example 1 minute, or maybe even 10 seconds.

Also, we might need to better document this issue (in addition to the warning in the log file). For non cluster case we can do away with lease time out and this for such cases indexing would not get paused upon restart post abrupt shutdown

Issue Links

• relates to: OAK-3923 Async indexing delayed by 30 minutes because stop order is incorrect
• relates to: OAK-3834 AsyncIndexUpdate default async lease timeout is (too?) long
• relates to: OAK-5159 Killing the process may stop async index update to to 30 minutes, for DocumentStore (MongoDB, RDB)