Project

General

Profile

Bug #6870

implement a request queue for remote OpenClient connections

Added by Constantin Asofiei over 1 year ago. Updated over 1 year ago.

Status:
New
Priority:
Normal
Assignee:
-
Target version:
-
Start date:
Due date:
% Done:

0%

billable:
No
vendor_id:
GCD
case_num:
version:

History

#2 Updated by Constantin Asofiei over 1 year ago

  • Subject changed from implement a request queue for remote appserver connections to implement a request queue for remote OpenClient connections

FWD has a request queue enabled for legacy REST/SOAP/web requests and for OpenClient connections from applications running in the same JVM as the FWD server.

When the application is executed standalone, each request is dispatched immediately to the FWD server. We need a request queue at the application level, which will start a number of workers (which should mimic the number of appserver agents available on the remote side), where the request is posted and delayed until the worker will finish executing the previous request.

#3 Updated by Greg Shah over 1 year ago

When you say "remote side", are you talking about a different FWD server (i.e. RUN ... ON in fwd_server_1 for a server handle that represents fwd_server_2)?

#4 Updated by Constantin Asofiei over 1 year ago

Greg Shah wrote:

When you say "remote side", are you talking about a different FWD server (i.e. RUN ... ON in fwd_server_1 for a server handle that represents fwd_server_2)?

I mean the application performing the request as a Open Client request, not via RUN statement.

#5 Updated by Constantin Asofiei over 1 year ago

Regarding the initial number of workers - once connection is established, it can interrogate the FWD server for the max agents value and default to the number of max agents.

#6 Updated by Greg Shah over 1 year ago

I'm surprised this is the case.

Are you saying that my understanding as documented in Remote Open Java Client Replacement is incorrect?

#7 Updated by Constantin Asofiei over 1 year ago

Greg Shah wrote:

I'm surprised this is the case.

Are you saying that my understanding as documented in Remote Open Java Client Replacement is incorrect?

Looks like it... I missed it, too.

But it won't be 'wrong' for long, the request queue is the only way to avoid 'no servers available', which can happen in OE.

OTOH, I need to check if FWD can 'bounce back' the appserver if all agents abend (i.e. via a STOP condition). There may be an issue here.

Also available in: Atom PDF