nv-l
[Top] [All Lists]

Re: APM

To: nv-l@lists.tivoli.com
Subject: Re: APM
From: "Patel, Shaileshbhai B" <shaileshbhai.patel@EDS.COM>
Date: Tue, 14 Dec 1999 16:42:01 -0500
Joel,

        Thanks for your input. Here is what I did to resolve the problem.

        ((@delta(.1.3.6.1.2.1.2.2.1.10.*)) *8*100/900) /
(.1.3.6.1.2.1.2.2.1.5.*)

 When you poll any mib variable using MLM (APM), it gets the real value of
the variable and not the delta value. So what I did is just poll inoctets
using APM(MLM) and at the same time use MIB browser to query the same mib
variable. Values were same.

After this, I poll the same mib variable but using APM, I got the delta
value by  (@delta(.1.3.6.1.2.1.2.2.1.10.*)). It gave the difference in value
over polling time which was 15 minutes.
Then we calculated bandwidth using above formula and results were matching
with results generated by xnmcollect application. Note that I have divided
the input octets(bit because I am multiplying it by 8) by 900 which is
polling time(15 minute) to convert the counter variable into gauge.

Here are my conclusions about this problem.

1). The expression is evaluated for mathematical operations. An expression
treats all integer, counter and gauge MIB variables as the type
     of output result. Therefore , when forming the MIB variable expression,
you need to know the type of MIB variable so that calculations and
     final results are consistent. Basically in the above case, what we did
is that, in the expression we converted the input octets rate into
     bits/sec and then divide it by speed which is also in bits/sec. Which
is gauge/gauge.

2). MLM polls the real value of the MIB variable.


Thanks,


> -----Original Message-----
> From: Joel A. Gerber [SMTP:joel.gerber@USAA.COM]
> Sent: Tuesday, December 14, 1999 4:11 PM
> To:   NV-L@UCSBVM.UCSB.EDU
> Subject:      Re: APM
>
> You may want to closely check the values that MLM is returning for the MIB
> objects.  Years ago we attempted to do MLM data collection and
> thresholding,
> but found that MLM does not apply the same delta calculations to COUNTER
> MIBs that NetView does.  I would recommend setting up a simple collection
> on
> a MIB object (instead of an expression) from both NetView and MLM and
> compare the results.
>
> Joel Gerber - I/T Networking Professional - USAA Information Technology
> Co.
> - San Antonio, TX
> * (210)456-4231         * mailto:Joel.Gerber@USAA.com   "
> http://www.usaa.com
>
>         -----Original Message-----
>         From:   Patel, Shaileshbhai B [SMTP:shaileshbhai.patel@EDS.COM]
>         Sent:   Tuesday, December 07, 1999 13:56
>         To:     NV-L@UCSBVM.UCSB.EDU
>         Subject:        APM
>         Importance:     High
>
>         Hello Everybody,
>
>                 I have NetView server 5.1.2 running on Solaris 2.6 O/S.
> Here
> is what
>         I am trying to do using Agent Policy Manager.
>                 Using threshold and Collection policy, obtain bandwidth
> utilization
>         on CISCO 8540 switch/router. Following is the formula
>
>                 ((.1.3.6.1.2.1.2.2.1.10.*) *8 *100)/.1.3.6.1.2.1.2.2.1.5.*
>
>                 This is the router with full duplex configuration. I am
> getting
>         strange utilization results ( like 325%, 137% ), which is not
> true.
>                 If I use xnmcollect from NetView, results are OK. I have
> also used
>         @delta function eventhough by default all counter MIB variables
> are
>         converted into integer and when query these variables it should
> give
> you the
>         delta value.
>                 Any suggestions?
>
>
>         Thanks,
>
>
>         Shailesh Patel
>         Enterprise Management Services
>         Troy MI 48098
>         (248) 265-7573


<Prev in Thread] Current Thread [Next in Thread>
  • APM, Patel, Shaileshbhai B
    • Re: APM, Leslie Clark
    • Re: APM, Patel, Shaileshbhai B
    • Re: APM, Joel A. Gerber
    • Re: APM, Patel, Shaileshbhai B <=

Archive operated by Skills 1st Ltd

See also: The NetView Web