Re: [HAPI-devel] Catastrophic failure...

2012-12-03 Thread Jens Kristian Villadsen
Well, throwing an Error instead of an Exception will do exactly what you wan't - it will make HAPI's receiving thread to terminate and nothing will be shipped over the wire. Try it yourself, it should be fairly easy to test 2012/12/3 LDH > We are on 1.2, with plans to move to 2.0 in the next f

Re: [HAPI-devel] Catastrophic failure...

2012-12-03 Thread LDH
We are on 1.2, with plans to move to 2.0 in the next few months. - We really want to simply fail and return nothing back to the sender, i.e., reset the connection in effect. The sending system will retry, and we'll refail again until we regain a db connection. etc... - I am not sure about your

Re: [HAPI-devel] Catastrophic failure...

2012-12-03 Thread LDH
That is interesting. We'll try that option and report back. Thank you, *Laurent Hasson* Co-Founder, 360Fresh.com Cellphone: 646-283-2186 This email and any attachments thereto may contain private, con

Re: [HAPI-devel] Catastrophic failure...

2012-12-03 Thread Christian Ohr
Laurent, What version of HAPI are you using? Internals of the whole app framework changed quite a bit in 2.0. If I understood you correctly, you need to close the socket on an erroneous message and you are well aware of the fact that you expose yourself to DOS attacks by doing so - anyway. In gen

Re: [HAPI-devel] Catastrophic failure...

2012-12-03 Thread Jens Kristian Villadsen
from within the Application.processMessage() method, that is 2012/12/3 Jens Kristian Villadsen > throw an Error instead of an Exception - that makes the HAPI thread choke > and nothing will be returned and the entire framework still lives on. That > sounds like its a proper solution for you. >

Re: [HAPI-devel] Catastrophic failure...

2012-12-03 Thread Jens Kristian Villadsen
throw an Error instead of an Exception - that makes the HAPI thread choke and nothing will be returned and the entire framework still lives on. That sounds like its a proper solution for you. 2012/12/2 LDH > > I have performed some stress testing this weekend. One test is to kill our > database

[HAPI-devel] Catastrophic failure...

2012-12-02 Thread LDH
I have performed some stress testing this weekend. One test is to kill our database, which causes our HAPI code that ingests messages to fail. Our strategy so far was simply to return a NACK, but we discovered that the system that is sending us messages is taking ANY response as a success ind