heartbeat_stats bf value increase


#1

Hy, I’ve a 3 nodes cluster on GCE with local SSDs configured as specified in documentation.
Replication factor of 2.
The cluster is well working.
When I’m looking at logs, I extract the heartbeat_stats lines and I’m surprised to see the bf value increasing each time :

heartbeat_stats: bt 0 bf 5567716 nt 0 ni 0 nn 0 nnir 0 nal 0 sf1 0 sf2 0 sf3 0 sf4 0 sf5 0 sf6 0 mrf 0 eh 0 efd 0 efa 0 um 0 mcf 1 rc 1
heartbeat_stats: bt 0 bf 5567918 nt 0 ni 0 nn 0 nnir 0 nal 0 sf1 0 sf2 0 sf3 0 sf4 0 sf5 0 sf6 0 mrf 0 eh 0 efd 0 efa 0 um 0 mcf 1 rc 1
heartbeat_stats: bt 0 bf 5568124 nt 0 ni 0 nn 0 nnir 0 nal 0 sf1 0 sf2 0 sf3 0 sf4 0 sf5 0 sf6 0 mrf 0 eh 0 efd 0 efa 0 um 0 mcf 1 rc 1
heartbeat_stats: bt 0 bf 5568336 nt 0 ni 0 nn 0 nnir 0 nal 0 sf1 0 sf2 0 sf3 0 sf4 0 sf5 0 sf6 0 mrf 0 eh 0 efd 0 efa 0 um 0 mcf 1 rc 1


Is it a configuration problem, or is it normal ? I was expected a 0 value instead, regarding the documentation as bf is "Received heartbeat packet on the wrong socket File Descriptor".

Thanks a lot.

Emmanuel.

#2

Hello. While not directly harmful, we do track stats like “bf” to try to get insight into potential cluster trouble. The fact that the cluster seems to be working well is a good sign. What release are you running? Do all 3 nodes show increasing “bf” counts? Does the number increase steadily, or has it (or does it sometime) stop? Thanks.


#3

Hello Our 3 nodes are running the same version : “Aerospike Community Edition build 3.5.3”. As soon as a node is included in the cluster, the “bf” value start to grow and never stop, as if it was the number of heartbeats received during the logging period. Otherwise, the cluster is well working with a replication factor of 2. Thanks.


#4

Hello. We have reproduced this issue. While it is not a serious problem, but rather more of an architectural untidiness, we have added fixing it to our backlog.

Please let us know if you have any more observations about this problem or its effects. Thanks for your help!