Martin > I've been all over the manuals, and I don't find this anywhere.
The APPN Tutorial redbook, now called "Inside APPN and HPR - The Essential Guide to New SNA", the first hit when selecting "APPN" as the search word on the redbooks page, is the manual you appear to have missed. http://www.redbooks.ibm.com/abstracts/sg243669.html > It appears that certain VTAM links, CTCs and others, between LPARs within and without of the same sysplex, get assigned a default TGN of 21. 21 is the TG number proposed for assignment to an APPN adjacent link when it (a PU statement) is defined without a specifying a TG number (using the TGN operand) (necessarily with a value between 1 and 20) and no link connection has yet been established between the two involved nodes. If neither side has specified a TG number, 21 will be used. Note that a second link under otherwise the same circumstances will use TG number 22 and so on. > Why? I direct you to section 4.2.3 Transmission Groups in the referenced redbook, specifically Table 2 and the associated explanation. > Shouldn't this be documented? It is - and where it should be as indicated. > More importantly, how does this affect PATH definitions, which contain VRs, and ER/TGN combinations, plus the COS tables that point to VRs, and the logmodes that point to COS tables? It does not affect those definitions one jot - except as a harbinger of getting rid of them altogether! This an APPN "thing" and nothing to do with subarea "things" such as - shudder - PATH tables. > ... plus the COS tables that point to VRs, and the logmodes that point to COS tables? Also not at all for the same reason. However, logmodes and COS are involved but indirectly and not in any "you'd better make sure it all matches or nothing will work" way. APPN routes get selected based on a loose form of matching between the characteristics of the transmission group as best defined by use of the TGP operand of the PU statement defining the adjacent link station. The mode name names a mode table entry (logmode) which may also contain an APPNCOS operand naming an APPN-flavour COS table (automatically activated when your NODETYPE start option specifies NN or EN) which selects a route based on finding a best fit with all the transmission groups making up potential routes - massively more dynamic than rigid, not forgetting to shudder - PATH tables designed as they were for storage-constrained 3705s back in the late '70s. > Please excuse my ignorance on this. I haven't dealt with VTAM for over 20 years. How come you are using APPN without having had APPN education? Let me guess that it's in preparation for Enterprise Extender which thoroughly inadequate guidance as to how to go about it. > Now, I've specified all the (apparently) valid TGNs we had defined (not including TGN 21), and I find some logmodes that can no longer be used between networks, while other logmodes still work. The logmodes are available on all systems, but it appears that there are no routes available for the logmode-selected COS entries. I'm now shuddering to try to imagine what you have been trying to do!!! I guess the fundamental trap your lack of education has allowed you to tumble into is that a subarea TG and an APPN TG are animals belonging almost to different phyla. > Maybe I'm making this more difficult than necessary. Education would help - but, given I based my APPN class on that redbook, you may as well simply read - and absorb - what the redbook says. Then you can read and absorb a couple more redbooks to which I now direct you: "Subarea to APPN Migration: VTAM and APPN Implementation" http://www.redbooks.ibm.com/abstracts/sg244656.html and "Subarea to APPN Migration: HPR and DLUR Implementation" http://www.redbooks.ibm.com/abstracts/sg245204.html Note that, if you are preparing for Enterprise Extender, you will also need to "get your head around" the HPR APPN extension. Then you can reinforce this by reading through the Network Implementation Guide for wherever APPN is mentioned. > The logmode entries worked previously, and TGN 21 was never included in the path definitions. Is there some default that allows any and all TGNs to be used by a given COS? "Education, education, education" a favourite expression of Tony Blair - of whom we have not heard much lately. In order to anticipate your next question - or the next question to come to any idly reading this without having your urgent need to read through the redbooks - TG numbers 1 to 20 are reserved so that they can be defined with the TGN operand of PU statements defining "APPN" adjacent link stations. This permits an APPN transmission group to be defined without any actual contact having been made. This then allows for such a link to be established dynamically when a session is requested where the APPN COS mechanism sets up a route which includes that transmission group. And if the defined numbers are different? Well, that's where this post response stops - because my dinner is ready - and redbooks are going to have to be downloaded! Chris Mason On Wed, 3 Sep 2008 10:23:38 -0500, Martin Kline <[EMAIL PROTECTED]> wrote: > copied above ---------------------------------------------------------------------- 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