Pat, Thanks for the update.
You are right on the issue here I'm happy to say. It may be that enough people pointed out that this fixation on subnet routes was ridiculous and the dam gave way. I've just now checked my mail from 2001 on this topic and I see I made as strong a case as I could through the most formal channel I had at the time. Maybe this contributed to the storm surge. Reading this up now that you've pointed this parameter out - with similar, and confusing in their excess, values for the SUBNET parameter - it seems CS IP development have completely reversed their position on advertising VIPA host routes. Indeed I too have "seen" a packet entering a sysplex on the promise of a subnet advertisement and being bounced between LPARs until the TTL expired. Each LPAR promised to one of its neighbours that the destination was covered by an equivalent subnet advertisement. That's what you experienced as the "perpetual routing". I can say I "saw" this effect since it's one of the functions of that package I keep mentioning to you that, based on routing tables, the progress of a packet from source to destination is fully explained or, as in this case, from source to the node where the TTL expires. Fortunately I foresaw the need for the TTL as the third parameter (in addition to source and destination IP address) or the routing simulation would still be running :-) Back in 2001 I did mention this to someone close to CS IP "central" and he said that they knew about the problem and some customers knew about the problem and they didn't care. I'll send you the full description privately although I expect it aligns precisely with what you also observed. Actually what I suggested aligns rather closely with what we now have with Advertise_VIPA_Routes - but it was bound to, being so obvious. Chris Mason ----- Original Message ----- From: "Patrick O'Keefe" <[EMAIL PROTECTED]> Newsgroups: bit.listserv.ibm-main To: <[email protected]> Sent: Thursday, 23 February, 2006 10:55 PM Subject: Re: TCP/IP over Cisco router CIP > On Thu, 23 Feb 2006 14:45:29 +0100, Chris Mason <[EMAIL PROTECTED]> > wrote: > > >... > >There was an issue with VIPAs and OMPROUTE OSPF dynamic routing. Ideally > it > >should be possible to advertise a VIPA using a "host", single (the VIPA > >itself) address rather than a subnet range. When I examined setting this > up > >in 2001, the last time I consulted on this topic, I seem to recall this > >ideal was not possible. Has anything changed or is it still necessary to > >assign, say, 4 addresses to each dynamic VIPA in order to advertise > >availability of individual dynamic VIPAs outside the LPAR? Of course, the > >waste may not really be a problem when assigning addresses from, say, > >network 10 in an intranet. > >... > > Chris, > > If I understand the issue here I asked that same kind of question 4 years > ago or so when setting up an OSPF config. I think I was told by Mike Fox > (OMPROUTE developer) that OSPF works on subnets and there wasn't any way > to get away from subnet advertisements. (At least that's what I thought > he said.) Since then a parm Advertise_VIPA_Routes has appeared, and it > can be set to HOST_ONLY. If you put that parm on all OMPROUTE's VIPA defs > the VIPA subnet won't be advertised. > > I sure wish I had that at my last shop. All our VIPAs were in the same > subnet spread across 7 LPARs. If a dynamic VIPA was not currently defined > and some external host tried to find it the packet would be perpetually > routed from host to host. (I don't have that addr, but he has that > subnet; try him.) :-( > > Pat O'Keefe ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
