nv-l
[Top] [All Lists]

RE: DPU Nodes reported Down/Up by Netmon

To: nv-l@lists.tivoli.com
Subject: RE: DPU Nodes reported Down/Up by Netmon
From: "Leslie Clark" <lclark@us.ibm.com>
Date: Mon, 17 Sep 2001 07:14:55 -0400
I cannot tell you from here. I can only give you information on how
Netview works. It is telling you something about your network, but
you will have to puzzle out what that is.

You might try looking at /usr/OV/log/netmon.trace after turning
tracing on (netmon -M -1). There may be some clues in there.

Santosh, do you subscribe to the listserver? It would be best if
we corresponded via the listserver, since others may be able to
help. I see listserver notes just as quickly as I see direct notes.

Cordially,

Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking
Detroit



                                                                                
                        
                    z_esm_karekars                                              
                        
                    <z_esm_karekars@bharatpetr       To:     Leslie 
Clark/Southfield/IBM@IBMUS          
                    oleum.com>                       cc:                        
                        
                                                     Subject:     RE: [NV-L] 
DPU Nodes reported Down/Up 
                    09/17/01 06:58 AM                 by Netmon                 
                        
                                                                                
                        
                                                                                
                        
                                                                                
                        



Hi Les,


I had implemented this the same way U had mentioned my timeout is 80 sec.
and retry is 8, but still some nodes responds in trapd.log as Down and
within 2-5 sec. it will report again Up by Netmon. How can get rid of this.


Please reply ASAP if possible.


Regards,
Santosh


     ----------
     From:   Leslie Clark[SMTP:lclark@us.ibm.com]
     Sent:   Monday, September 17, 2001 4:24 PM
     To:     z_esm_karekars; NV-L@tkg.com
     Subject:        RE: [NV-L] DPU Nodes reported Down/Up by Netmon


     The default is to ping. They are only polled via SNMP under these
     circumstances:
     1) you put them in the seedfile with $
     2) you put a P flag in the oid_to_type file for that snmp sysobject id

     3) the node is a router with an unnumbered serial interface that
     connects
     to another router with an unnumbered serial interface.


     Therefore, if, as you say, the nodes are being polled via SNMP, and
     they
     are not snmp enabled, it can only be because you configured netview to
     do
     this by method 1 above. There should be no need to force it to go the
     other
     way, and so there is no method to do that, as far as I know. You can
     tell
     from
     the output of demandpoll which way it is using to obtain interface
     status
     info
     (as opposed to configuration info).


     Cordially,


     Leslie A. Clark
     IBM Global Services - Systems Mgmt & Networking
     Detroit







                         z_esm_karekars

                         <z_esm_karekars@bharatpetr       To:     Leslie
     Clark/Southfield/IBM@IBMUS
                         oleum.com>                       cc:

                                                          Subject:     RE:
     [NV-L] DPU Nodes reported Down/Up
                         09/17/01 03:19 AM                 by Netmon









     Hi Les,





     Sorry to bother U again, Can U pls mail me the steps how can I
     configure
     and IP address to be polled for status through Ping and not through
     SNMP
     Status Poll.





     Regards,
     Santosh





          ----------
          From:   Leslie Clark[SMTP:lclark@us.ibm.com]
          Sent:   Friday, September 14, 2001 9:54 PM
          To:     z_esm_karekars
          Subject:        RE: [NV-L] DPU Nodes reported Down/Up by Netmon





          Absolutely. If they are not snmp-enabled, then Netview cannot
     poll
          them
          via SNMP. If I gave you the impression that they could, I
     apoligize.
          Only
          snmp-capable things can be polled via SNMP. Remove the $ entries
          in the seedfile, then stop and restart netmon. It should change
          methods
          on the next round, but if it does not appear to take effect right

          away,
          select a bunch of them and demandpoll them (from a smartet,
     perhaps).





          Cordially,





          Leslie A. Clark
          IBM Global Services - Systems Mgmt & Networking
          (248) 552-4968 Voicemail, Fax, Pager










                              z_esm_karekars


                              <z_esm_karekars@bharatpetr       To:
     Leslie
          Clark/Southfield/IBM@IBMUS
                              oleum.com>                       cc:


                                                               Subject:
     RE:
          [NV-L] DPU Nodes reported Down/Up
                              09/14/01 09:28 AM                 by Netmon












          Hi Les,








          As U had mentioned to enter $ in front of the IP Address of the
     DPU's
          Do U
          suggest this as they are not SNMP complient ?, but the help in
     seed
          file is
          not clear as follows. Should I try this to enable ICMP ping for
     these
          DPU's
          in place of SNMP status poll ?








          Please advice ASAP








          Regards,
          Santosh








          #    SNMP Status Support
          #
          #      Nodes that have serial interface cards will automatically
     use
          SNMP
          #      to query for ifAdminStatus and ifOperStatus for ALL
     interfaces
          on
          #      that node.  Specify other nodes you would like to use SNMP
     to
          query
          #      for status instead of ICMP pings here.
          #
          #      Use the prefix '$', and use the same syntax for ranges and

          wildcards





          #      described above for positive seed entries.  For instance,
          #
          #        $1.1.2-10.221-230
          #








               ----------
               From:   Leslie Clark[SMTP:lclark@US.IBM.COM]
               Reply To:       IBM NetView Discussion
               Sent:   Friday, September 14, 2001 6:36 PM
               To:     IBM NetView Discussion
               Subject:        RE: [NV-L] DPU Nodes reported Down/Up by
     Netmon








               Are the events faster than your polling cycle (eg 5
     minutes)?
               If the up/down events are coming faster than your polling
     cycle
          then
               there are two possibilities:
               1) They are sending traps (link up, link down) and Netview
     is
          reacting
               to
               them,
               and frequently catching them in an actual down state. Not
     likely,
          but
               possible
               2) Router Fault Isolator is running, and misbehaving due to
     the
          nature
               of
               you
               network. Perhaps there is a route to that subnet that
     Netview
          cannot
               see.
               In
               that case, you can turn off RFI by adding '-K 0' as a
     parameter
          in the





               netmon.lrf file. The procedure is described in the 6.0.1
     release
               notes.








               If the events are the same frequency as your polling cycle,
     eg 5
               minutes
               apart, then it is either a timeout, or it has to do with the

               difference
               between
               status via ping or status via snmpget. If these devices are
     not
          snmp
               capable,
               be sure you are not polling them via SNMP. For instance, if
     you
               entered
               them in the seedfile with $, or put a P in the oid_to_type
     file
          for
               them,
               then
               the demandpoll (or regular netmon poll) will turn them red,
     but a


               test...ping will
               turn them green.








               I know none of this sounds exactly like your situation, from
     the
               infomation
               you have given us, but perhaps these clues will help you
     figure
          out
               what
               is going on.








               Cordially,








               Leslie A. Clark
               IBM Global Services - Systems Mgmt & Networking
               Detroit













                                   z_esm_karekars





                                   <z_esm_karekars@bharatpetr       To:
          "'IBM
               NetView Discussion'" <nv-l@tkg.com>
                                   oleum.com>                       cc:





                                   Sent by:
     Subject:
          RE:
               [NV-L] DPU Nodes reported Down/Up
                                   owner-nv-l@tkg.com                by
     Netmon









                                   09/14/01 07:56 AM





                                   Please respond to IBM





                                   NetView Discussion














               Hi,











               I tried setting the SNMP polling interval and timeout as
     high as
               possible
               but still trapd.log gets updated with Netmon events for
     these
          DPU's at
               very
               fast Down/Up events. Where else could be the problem.











               Regards,
               Santosh











                    ----------
                    From:   Jane Curry[SMTP:jane.curry@skills-1st.co.uk]
                    Reply To:       IBM NetView Discussion
                    Sent:   Thursday, September 13, 2001 3:26 PM
                    To:     IBM NetView Discussion
                    Subject:        Re: [NV-L] DPU Nodes reported Down/Up
     by
          Netmon











                    Put an entry for these devices in your ovsnmp.conf
     (Options
          ->
               SNMP
                    Configuration).  I would suggest increasing both the
     timeout


               interval
                    and also, perhaps, the retries if the lines are a
     loss-y.











                    Cheers,





                    Jane
                    z_esm_karekars wrote:














                         Hi,











                         I have Netview 6.0.1 on AIX 4.3.3











                         We have almost 180 DPU (SNMP Not supported)
     locations
          on
               slow
                         link in our Wide Network. It is reported in
     trapd.log
          that
               very
                         frequently they go Down and come Up so fast (which

          could be
                         because of slow link response). If I do a query to

          database
               for
                         these DPU's Up/Down status, it becomes a huge and
     non
               relevent
                         data.











                         Can I have some solution for this so that the
     genuine
          DPU
               Down
                         and Up will be reported to trapd.log











                         Regards,
                         Santosh











                    --
                    Tivoli Certified Enterprise Consultant & Instructor
                    Skills 1st Limited, 2 Cedar Chase, Taplow, Bucks, SL6
     0EU,
          UK
                    Tel: +44 (0)1628 782565
                    Copyright (c) 2001 Jane Curry
     <jane.curry@skills-1st.co.uk>.
          All





                    rights reserved.


















     _________________________________________________________________________







               NV-L List information and Archives: http://www.tkg.com/nv-l


<Prev in Thread] Current Thread [Next in Thread>

Archive operated by Skills 1st Ltd

See also: The NetView Web