Understanding the gateway.recover_after_time setting

OpenSearch
,
1

Versions (OpenSearch): 3.5.0

Describe the issue:

If both gateway.recover_after_data_nodes and gateway.expected_data_nodes are set and contain different values, which of those settings will be used by gateway.recover_after_time to determine if enough data nodes are present to start the recovery?

While the documentation describes the gateway.recover_after_time setting as

The maximum amount of time to wait until recovery if the expected data node count hasn’t been reached.

, it is unclear to me if the “expected data node count” will change to the value supplied by gateway.recover_after_data_nodes if both gateway.recover_after_data_nodes and gateway.expected_data_nodes are set.

2

@bennethinz recover_after_time is evaluated against expected_data_nodes , not recover_after_data_nodes . The “expected data node count” in the docs refers specifically to the gateway.expected_data_nodes value. The recover_after_data_nodes setting just controls the earlier hard gate before any of timer logic kicks in.

Hope this helps

3

Thank you for the quick answer!

1 Like