Re: [Fwd: Pls post this to forces WG list.]

Robert Haas <rha@zurich.ibm.com> Tue, 09 September 2008 12:52 UTC

Message-Id: <TUE.9.SEP.2008.145237.0200.>
Date: Tue, 9 Sep 2008 14:52:37 +0200
From: Robert Haas <rha@zurich.ibm.com>
Subject: Re: [Fwd: Pls post this to forces WG list.]
Comments: To: "Bert Wijnen (IETF)" <bertietf@bwijnen.net>
Comments: cc: "Romascanu, Dan (Dan)" <dromasca@avaya.com>
MIME-Version: 1.0
Content-Type: multipart/alternative; boundary="=_alternative 0046BBEDC12574BF_="

Thanks Bert, I made the changes as suggested, the updated draft is:

http://www.ietf.org/internet-drafts/draft-ietf-forces-mib-09.txt

My summary of the changes is copied below (See Section 10 in the draft).
Regards,
-Robert

Changes from draft-ietf-forces-mib-08:

   o  Changed the MIB objects forcesAssociationOtherMsgSent and
      forcesAssociationOtherMsgReceived to forcesAssociationOperMsgSent
      and forcesAssociationOperMsgReceived as they are not all other
      messages besides HB (comment from the General Area Review Team).

   o  Changed MIB counter objects forcesAssociationHBMsgSent,
      forcesAssociationHBMsgReceived, forcesAssociationOperMsgSent, and
      forcesAssociationOperMsgReceived from Counter32 to
      ZeroBasedCounter32 (comment from Bert Wijnen).  Adapted the
      paragraph about statistics counters in section "Associations kept
      in the MIB"

   o  Introduced a MIB object forcesAssociationCounterDiscontinuityTime,
      and added it to the forcesAssociationEntry as well as to the
      forcesAssociationEntryDownStats notification and the
      forcesStatsGroup compliance group.  Added text on discontinuity
      for all counter objects.

   o  Removed MIB counters from the forcesAssociationEntryUpStats
      notification, as passing now zero values is useless.



"Bert Wijnen \(IETF\)" <bertietf@bwijnen.net> wrote on 09/08/2008 10:39:12
PM:

> [image removed]
>
> Re: [Fwd: Pls post this to forces WG list.]
>
> Bert Wijnen \(IETF\)
>
> to:
>
> Robert Haas
>
> 09/08/2008 10:40 PM
>
> Cc:
>
> FORCES, "Romascanu, Dan \(Dan\)"
>
> From your description it sounds that indeed a ZeroBasedCounter32 would
be
> better. That way you indicate it starts at zero.
>
> And hearing that the same index (FEID+CEID) could come to life in the
> table after one has been deleted, that makes it as a discontinuity.
> How else would a NM appliation know that this is a different entry?
> And so a discontinuity timestamp is needed.
>
> If there are no other reason why a discontinuity can occur (I do not
know,
> because I know to little of forces), then it would be good to state so.
>
> Text on page 32 of
>    > http://www.wtcs.org/snmp4tpc/FILES/Reference/snmp%20counters.pdf
> is an example of a discontinuity counter.
>
> Bert
> ----- Original Message -----
> From: "Robert Haas" <rha@zurich.ibm.com>
> To: <bertietf@bwijnen.net>
> Sent: Monday, September 08, 2008 7:44 PM
> Subject: Re: [Fwd: Pls post this to forces WG list.]
>
>
> > Bert,
> > Indeed, I refrained as I saw that the use of ZeroBasedCounter32 was
often
> > discouraged ... But in the ForCES MIB, when the ForCES association
comes
> > up, a table row is created, and when the association goes down, the
table
> > row is destroyed. So it would be more natural if the counters in that
row
> > were zero-based (zeroed upon creation of the table row) instead of
> > starting at a random value. If you say it's an acceptable use case
then
> > I'll change to ZeroBasedCounter32.
> >
> > Note that an association is identified by its index which is the
> > concatenation of the FEID and CEID forming the association. So a new
> > association coming up between the same FEID and CEID would appear with
the
> > same index as a previous association that went down. Would this be
> > considered a discontinuity as the counters appear to be reset ? Apart
from
> > that, what else would cause a discontinuity ?
> > I'd use the text from page 32 in
> > http://www.wtcs.org/snmp4tpc/FILES/Reference/snmp%20counters.pdf to
> > address this.
> >
> > Thanks,
> > -Robert
> >
> >
> >
> >
> >
> > From:
> > Jamal Hadi Salim <hadi@znyx.com>
> > To:
> > FORCES@PEACH.EASE.LSOFT.COM
> > Date:
> > 09/08/2008 03:07 PM
> > Subject:
> > [Fwd: Pls post this to forces WG list.]
> >
> >
> >
> > Bert,
> > Forwarded to Robert and Forces list (sorry couldnt figure how to let
the
> > allow you to post without subscribing you)
> >
> > cheers,
> > jamal
> >
> > ----- Message from "Bert Wijnen \(IETF\)" <bertietf@bwijnen.net> on
Mon, 8
> > Sep 2008 11:17:42 +0200 -----
> > To:
> > dro@zurich.ibm.com, hadi@znyx.com
> > Subject:
> > Pls post this to forces WG list.
> > I am not on the list and do not want to subscribe either. So my
> > posting got bounced. Pls forward or post it for me.
> >
> > So if people want to comment/respond, they should explicitly
> > copy me.
> >
> > Bert Wijnen
> > ----- Original Message -----
> > From: "Bert Wijnen (IETF)" <bertietf@bwijnen.net>
> > To: <ietf@ietf.org>
> > Cc: <forces@peach.ease.lsoft.com>
> > Sent: Monday, September 08, 2008 10:54 AM
> > Subject: Re: Last Call: draft-ietf-forces-mib (ForCES MIB) to
> > ProposedStandard
> >
> >
> >>I sort of wonder if the Counter32 is the proper datatype for some
> >> of the counters. They sound more like ZeroBasedCounter32 to me.
> >>
> >> Further I do not see any text regarding possible discontinuities.
> >>
> >> Bert Wijnen
> >>
> >
> >
> >
> >
>
>