To: | nv-l@lists.us.ibm.com |
---|---|
Subject: | Re: [nv-l] Re: [nv-l] Discovering networks without switches but with DHCP Nodes |
From: | Leslie Clark <lclark@us.ibm.com> |
Date: | Wed, 12 Jan 2005 17:13:20 -0500 |
Delivery-date: | Wed, 12 Jan 2005 22:09:56 +0000 |
Envelope-to: | nv-l-archive@lists.skills-1st.co.uk |
In-reply-to: | <OF9AC0E08A.1A21964B-ON87256F86.0066404C-85256F87.0074CAD7@us.ibm.com> |
Reply-to: | nv-l@lists.us.ibm.com |
Sender: | owner-nv-l@lists.us.ibm.com |
About discovery of remote subnets where you don't have an snmp-enabled router: Netview would normally discover things from the arp cache of other things which are snmp-enabled. If you have nothing there, you will have to force the discovery. You can do this in a few ways. You can put the addresses, or names, in your seedfile. It will do a pretty good job. It may get some subnet masks wrong, but it will tell you about it (in netmon.trace) so you can delete them and rediscover them. You can also run some sort of a pinger to ping them. If Netview hears from them, it will add them if the seedfile allows it (eg there are no restrictions in it for those addresses). What you will get is a bunch of disconnected subnets, since there are no routers to connect the subnets. This does not affect Netview's ability to monitor their status. About DHCP on the Unix platforms: Netview on the Unix platforms does not handle these in any useful way. On the Windows platform, you can tell it in the seedfile that it is DHCP, and it will know what to do when it changes. On the Unix platforms, it is possible that it will figure it out based on changing name resolution, but not likely. What most people do is exclude the DHCP nodes from discovery, by specifying the addresses or address ranges in the seedfile with a '!" in front of them so they are excluded. Since they are probably desktop pc's, and they get turned off every afternoon, no one wants them in the map where all they do is turn red. In your case, since you want to monitor them, then you were better off on Windows. If I had to try to make the Linux version handle DHCP nodes, I would try: 1) make sure 'Node Down Delete' interval is short for those address ranges, like 1 day. This is under Options...SNMP Config' 2) make sure name resolution is dynamic - that is, when a dhcp assignment changes, the name changes in your DNS 3) the configuration interval is short for those address ranges, like 4 hours. This is also under Options...SNMP config. It is possible that normal name updating will handle the deleting of the old addresses. Something else I have tried when dealing with DHCP nodes is to create a Netview Ruleset that listens for Interface Added events, and invoke a script. That script can figure out if this added interface is a DHCP one, and if there is already a down interface on that node, and then the script can delete the old interface. That would be an advanced topic, but it is something to think about if you can't come up with anything else. Cordially, Leslie A. Clark IBM Global Services - Systems Mgmt & Networking (248) 552-4968 Voicemail, Fax, Pager
Camron, Please don't send me personal notes. My replies have to be to the list. Do you have SNMP enabled anywhere? NetView is an SNMP-manager and automatic discovery is going to depend on reading the ARP caches of the devices you have discovered, using SNMP, to find new devices. If you have no SNMP anywhere, then you are going to have to populate your database manually using the loadhosts command. But I thought you said that when you manually pinged devices, then netmon added them to the map correctly. Are the subnet masks correct or not? If not, then once again you will have to use loadhosts to add new nodes, and you may very well have to delete the "bad" once you have and load them afresh correctly, since without SNMP, netmon would have no way to confirm that what he guessed as the subnet mask was correct. You can put not non-SNMP nodes in the seed file, but as the doc explains, there is risk that they will be discovered with he wrong subnet mask since netmon will have to guess at it. The issue of DHCP is murkier, and not being the netmon expert, I'm not certain what to say. Perhaps someone else will respond. But if your DHCP nodes don't have SNMP, then I think we are going to have a hard time identifying them as DHCP. If you don't get the response your are looking for here, then I suggest a call to IBM Support, so that you can get personalized help. James Shanks Level 3 Support for Tivoli NetView for UNIX and Windows Tivoli Software / IBM Software Group Camron R <gundulf@walla.com>
Thanks, James. But I still have questions: 1) Can I put not SNMP supported nodes on the seed file? 2) Can I use the same flagging of DHCP nodes from the seed file on Windows version on the Linux version?? <jshanks@us.ibm.com> Re: [nv-l] Discovering networks without switches but with DHCP Nodes
|
<Prev in Thread] | Current Thread | [Next in Thread> |
---|---|---|
|
Previous by Date: | Re: [nv-l] Re: [nv-l] Discovering networks without switches but with DHCP Nodes, James Shanks |
---|---|
Next by Date: | Re: [nv-l] NV tuning for Data collection, Leslie Clark |
Previous by Thread: | Re: [nv-l] Re: [nv-l] Discovering networks without switches but with DHCP Nodes, James Shanks |
Next by Thread: | [nv-l] Event Attributes, Mildeberger, Thorsten |
Indexes: | [Date] [Thread] [Top] [All Lists] |
Archive operated by Skills 1st Ltd
See also: The NetView Web