MySQL Forums
Forum List  »  NDB clusters

Re: ConnectCheckIntervalDelay vs. HeartbeatIntervalDbDb
Posted by: Jonathon Coombes
Date: November 09, 2016 10:03PM

My understanding is the ConnectCheckIntervalDelay option is set to 0 by default meaning it is disabled. The HeartbeatIntervalDbDb cannot be disabled as such.

The heartbeat method is the primary one used for Cluster health.
When the heartbeat detection fails/faults in some way, then a NODE_PING_CONF is sent between nodes to determine if they are running.
Latency between nodes will determine if NODE_PING_CONF is returned from a peer node or not.

If latency is a known issue in certain setups (such as re-routing, failover etc) the connectivity check can prepare for the FAIL_REP signals that will flood in.
These connections will be marked as suspect connections and FAIL_REP signals originating from that node will be ignored for the ConnectCheckIntervalDelay period.

At least that is how I understand it all :)

Options: ReplyQuote


Subject
Views
Written By
Posted
Re: ConnectCheckIntervalDelay vs. HeartbeatIntervalDbDb
301
November 09, 2016 10:03PM


Sorry, you can't reply to this topic. It has been closed.

Content reproduced on this site is the property of the respective copyright holders. It is not reviewed in advance by Oracle and does not necessarily represent the opinion of Oracle or any other party.