Imported from previous forum
We are receiving test requests on this tag & wish to know if the fix engine ( quickfix ) should respond directly ( like a ping ) or needs to go through the queue ( like a ‘normal’ ) request, for the responding heartbeat to be made. Initially in high volume situations, we had not been able to respond in time, although subsequently we simply did not seem to respond at all. Any information appreciated.
[ original email was from SENTHIL SUBRAMANIAM - sensubra@lehman.com ]
> We are receiving test requests on this tag & wish to know if the fix
engine ( quickfix ) should respond directly ( like a ping ) or needs to
go through the queue ( like a ‘normal’ ) request, for the responding
heartbeat to be made. Initially in high volume situations, we had not
been able to respond in time, although subsequently we simply did not
seem to respond at all. Any information appreciated.
On high volume situation where there is a message flow in both directions within the heartbeat intervals, either of the fix engines should not be sending heartbeat msgs, btw test request should be used if there is no repsonse to heartbeat.
Even if so, response to test request, ie heartbeat msg should go thru normal application processing queue and receiving side should reset the heart beat timer even if it receives any message from other side.