nv-l
[Top] [All Lists]

Re: Reliability of NV Status Polling

To: nv-l@lists.tivoli.com
Subject: Re: Reliability of NV Status Polling
From: James_Shanks@TIVOLI.COM
Date: Tue, 25 Aug 1998 10:34:14 -0400
Reply-to: Discussion of IBM NetView and POLYCENTER Manager on NetView et alia <NV-L@UCSBVM.UCSB.EDU>
Sender: Discussion of IBM NetView and POLYCENTER Manager on NetView et alia <NV-L@UCSBVM.UCSB.EDU>
You and your customer are comparing apples to oranges, which is why you
think that NetView's polling is unreliable.

NetView is monitoring your network availability, not  an individual
machine's health.  It is perhaps unfortunate that the word "down" was
chosen to describe the condition when the target machine does not answer,
but that is the shorthand that is everyday use.  It does not matter whether
the source of the problem is that the network is too congested deliver the
ping, or that a router is so overburdened that SNMP requests are being
lost, or that the machine is looping and will not respond.  All of these
amount to the same thing.  That machine is not able to fully participate in
the network and is there marked as "Critical", which we colloquially call
"down".  That is what NetView, and any other SNMP-based monitor will tell
you.  When you want to know why it is "down' you have to use other tools to
find out why -- such as collect traffic statistics from your router and so
on.

James Shanks
Tivoli (NetView for UNIX) L3 Support



"Jeremy W." <jw@UK.IBMMAIL.COM> on 08/25/98 04:16:20 AM

Please respond to Discussion of IBM NetView and POLYCENTER Manager on
      NetView et alia <NV-L@UCSBVM.UCSB.EDU>

To:   NV-L@UCSBVM.UCSB.EDU
cc:    (bcc: James Shanks)
Subject:  Reliability of NV Status Polling





===============================================================

All,

We use both NV/6000 4.1 and NV5.0a/NT. Our customer is very anxious to
get us to agree that NetView status polling may be unreliable, and that
interfaces may appear to be down when in fact, all that is happening is
that the TCP/IP protocol on the machine being polled is TOO BUSY to
answer the ICMP request!!!

Is this in any way possible on TCP/IP for NT or AIX?

If the machine is TOO BUSY to answer a ping, is it not also TOO BUSY to
answer request for resources (i.e. it is effectively down anyway)?

Is it possible that some WAN router between networks could be TOO BUSY
to handle an ICMP packet? Would it not just keep trying?

Any help from you network techies would be very useful!

Thanks,

Jeremy.

---------------------------------------------------------------
 Jeremy Worrell (CS), IBM UK Ltd.            +44 (1926) 462464
 NT Support                                  jw@uk.ibmmail.com
---------------------------------------------------------------

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

Archive operated by Skills 1st Ltd

See also: The NetView Web