Closed aquette closed 6 years ago
Note on the [DO NOT MERGE YET] I want to make more tests on the case where fty-info fails to answer (at all or in a timely fashion), to ensure reliability and resiliency of behavior
It seems to me that HW_CAP is quite important message. But it is sent through same mlm instance that is used for listening to stream. It could be possible on busy systems (especially after reboot) that the stream channel would be busy and message from stream will be intercepted instead of reply, leading to a problem. Could you please separate mlm instances for messages and streams?
@EldoreiJK not sure it's worth, there is already an async retry mechanism through zloop timer. If you really feel it's needed, I'll do, but I really doubt it will make a difference
Solution: Implement it, with more tests
Signed-off-by: Arnaud Quette ArnaudQuette@Eaton.com