Mary I've had another look at this and I see I missed something that is possibly important about what you said.
> We have set up SNMP on the HMC with community names and IP addresses. What this is all about is allowing *access* for SNMP *requests*, GET, GETNEXT and SET, to the MIB managed by the SNMP agent in the HMC IP node so that *solicited* information can be returned to the SNMP manager. It does nothing for TRAPs, *unsolicited* information passed to conceptually an SNMP manager, which is what you appear to want. Since (a) I know almost nothing about the HMC, not being a "coalface" worker like most of the denizens of the list, and (b) I have no idea what your level of skill is regarding configuring IP nodes, let alone IP nodes performing the role of an SNMP agent on one side, the HMC, and SNMP managers on the other side, whatever this "automation point" is, I must assume the worst, as it were. The first thing you need to do is make sure that the HMC node is properly configured as an IP node. When I was having to set up "workstations"/"servers" for hands-on education purposes which I had to do for a dozen or so machines prior to every class, there were always 5 - the number of fingers on a hand! - elements to ensure were configured. It was particularly important that they were "ticked off" since this dated back to the days when OS/2 PCs had to be configured with line commands! - The node itself - (1) a name - The interface to the LAN - (2) an IP address and (3) a mask which when "anded" with the IP address defined an address range which should apply to the IP addresses of each interface connected to the same LAN, aka, the *subnet* mask, which enables the Address Resolution Protocol (ARP) and routing generally to operate correctly - Routing - (4) the IP address of the "adjacent" router interface which allows access to IP nodes other than those with interfaces directly connected to the same LAN - Name/address translation - (5) the IP address - often two for backup - of a server node which can translate names to IP addresses and IP addresses to names If you can PING from the platform supporting the "automation point server" to the HMC IP address, you're making progress. You've either done the above or you can ensure it is done and then you can proceed to configuring the SNMP agent in the HMC. This is documented in Chapter 6 of S/390 Application Programming Interfaces, not a name that you would immediately associate with the HMC - which may be why it doesn't feature on the HMC "bookshelf". I checked with my notes on what is needed in order to set up a basic SNMP agent - and SNMP manager - from long atrophied presentation on "TCP/IP for MVS". It should, like the basic IP node configuration above, ensure that the required parameters are covered. In support of authorising SNMP requests, GET, GETNEXT and SET, it is necessary to specify one or more community names and one or more associated sets of IP addresses - and, with some implementations, object identifier ranges within the MIB and whether, in addition to the "get"-type requests, SET requests are permitted - unless the default "public" for access of any sort from any IP address is accepted. It's probably best not to specify any complicated limitations here until you are sure the rest of the configuration is working. I expect this is what you meant by "community names and IP addresses". In order to provide a port to which SNMP requests can be sent, a port number must be specified unless the default 161, the designated "well-known" port for this function, is acceptable, which it nearly always will be. Depending in detail on the implementation, certain MIB variables in the "sys" group should be specified. I see that, in the case of the HMC, these are sysDescription, sysContact, sysName and sysLocation. Finally, there needs to be some specification of the destination of TRAPs - which I guess is what this thread is all about. In my last post I mentioned that this "specification of the destination of TRAPs" in the case of the HMC is the "Event notification information" box. While the actual IP addresses can be specified here, note that the port number used cannot be changed and is 162, the designated "well-known" port for this function. I also know nothing about this "automation point server" but it will not receive any TRAPs unless whatever the platform for the product is has some application using the API of the IP logic which "opens" port 162 to receive IP packets. Whether the TRAPs use UDP or TCP transport is not specified in the HMC manual. Traditionally - rather obviously given the nature of the "application" - it is UDP. I see access to the CA documentation is blocked and requires - as one contributor in this list rather aptly put it - that you promise them your first- born. No doubt this explains why you need someone who has actually set up CA "automation point" since there's just about no chance you can get help from people who simply fancy themselves as able to extract gold from the usual dross found in non-IBM manuals. Chris Mason On Fri, 24 Jun 2011 06:56:12 -0400, Mary Elwood <mary_elw...@navyfederal.org> wrote: >Good Morning, > >We are trying to get the HMC to send hardware messages, state change >messges, etc., to automation point. We have set up SNMP on the HMC with >community names and IP addresses. But we don't believe the API is sending >the message to the automation point server. > >Has anyone ever setup automation point from CA? > >Any and all help is greatly appreciated. > >Thank you, > >Mary ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
