Does fault detection process work sequentially?


(Yu Watanabe) #1

Hi.

I was reading the document and was bit unclear with the fault detection process.

Are the zen ping sent sequentially each time? Or , is it asynchronously regardless of response time of each ping?


(Alexander Reelsen) #2

Do you mean the pings to reach the nodes? This is done asynchronously there is no wait period per host or something. The nodes are not pinged sequentially. Otherwise there would not be any interval guarantees.


(Yu Watanabe) #3

@spinscale

Thanks for the reply.

I meant by on each host .

So, if you set each parameter as below,

ping_interval : 1s
ping_timeout : 30s
ping_retries : 3

Suppose you are sending ping to NodeB from NodeA and if timeout occurs on first ping check , does elasticsearch send second ping to NodeB after 30 s ?

e.g

1st try : Node A ---------> Node B (NG)
2nd try : Node A ---------> Node B <- After 30 sedonds


(system) #4

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.