Bug 1284 - Semantics of Connection_Timeout policy with multiple NIC's
Summary: Semantics of Connection_Timeout policy with multiple NIC's
Status: RESOLVED FIXED
Alias: None
Product: TAO
Classification: Unclassified
Component: ORB (show other bugs)
Version: 1.2.3
Hardware: All All
: P3 normal
Assignee: Nanbor Wang
URL:
Depends on:
Blocks: 1277
  Show dependency tree
 
Reported: 2002-08-15 14:06 CDT by Nanbor Wang
Modified: 2002-11-18 21:35 CST (History)
0 users

See Also:

Attachments
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Nanbor Wang 2002-08-15 14:06:27 CDT 
The semantics of Connection_Timeout policy is more 
like "RelativeRoundtripTimeoutPolicy" which prevents that being used for lots 
of nice use cases. 

Imagine a use case where there are two paths to the server, each through 
different NIC's. If the first NIC is down on the server or is bad, the clien t 
wouldnt be able to reach the first NIC. In such cases the ORB should try the 
second NIC. But the adoption of the semantics of "RelativeRound.." policy 
doesnt handle this scenario well enough. The ORB throws an exception 
CORBA::TIMEOUT back to the client when the first path is found invalid before 
the first invocation. 

This needs to be fixed and is a blocker for 1.3
Comment 1 Carlos O'Ryan 2002-08-15 17:07:10 CDT 
Blocks the 1.3 release, without this functionality fail-over is simply impossible.
Comment 2 Nanbor Wang 2002-08-23 17:56:42 CDT 
Accepting the bug
Comment 3 Nanbor Wang 2002-11-18 21:35:53 CST 
Fixed! Please see

Thu Oct 31 14:48:39 2002  Balachandran Natarajan  <bala@isis-
server.isis.vanderbilt.edu>

for details