ken@cs.cornell.edu (Ken Birman) (05/15/91)
> From: Flavio De Paoli <DEPAOLI%IMICEFR.BITNET@CORNELLC.cit.cornell.edu> > Subject: ISIS > To: Kenneth Birman <birman@cs.cornell.edu> > Status: R > Hi Ken, I have a problem with ISIS. > It seems that the forward mechanism does not work well. > Here is the code I used: > client: address send_address; > ... > cbcast(&send_address, ENTRY, ...); > server1: entry_routine(message* msg) { > address forward_address; > ... > forward (msg, &forward_id, ENTRY, msg); > server2: entry_routine(message* msg) { > ... > reply (msg, ...) > The point is that the client never get the reply. > I checked the sender's address with the pmsg function: in both server1 and > server2 msg has the same sender. Every think seems to be correct. > Why the reply fails? > Thank you for helping. > Flavio We are aware that this mechanism is broken in ISIS V2.1, but unfortunately the fix involves quite a few minor changes to the system. I do have this working in V3.0 and when V2.2 comes out, I plan to include the fix. Sorry that I can't just send a patch script, but in this case I don't have one. Sorry about this! It actually caught us by surprise too, some time ago, and I think I listed it in the posting to comp.sys.isis on "known bugs" As an interim solution you may need to program a mechanism in which you actually reply to the original sender and the sender re-issues the request to the new destination. Another problem, for those who might run into it, is that the token tool doesn't handle the failure of the token holder and yet a third problem is that certain failure sequences can cause a group to go away without pg_detect_failure noticing. All of these (any others, too) are fixed in V3.0 and will eventually be fixed in V2.2. We still have no specific date in mind for V2.2, though. -- Kenneth P. Birman E-mail: ken@cs.cornell.edu 4105 Upson Hall, Dept. of Computer Science TEL: 607 255-9199 (office) Cornell University Ithaca, NY 14853 (USA) FAX: 607 255-4428