Mukesh.Gupta | 2 Jun 04:49 2004
Picon

Agenda for IETF 60th

Me and Radia will be collecting the agenda items for the
VRRP WG meeting in IETF 60th, San Diego.

If you have something to discuss, please send it to Me and
Radia.

If we don't receive any agenda items, we might decide not to
meet.

Regards
Mukesh

_______________________________________________
vrrp mailing list
vrrp <at> ietf.org
https://www1.ietf.org/mailman/listinfo/vrrp

Jahanzeb Faizan | 9 Jun 15:58 2004

failure model

 
Hi,
 
I want to know if there is any failure model defined for router failures in vrrp. Also I want to ask how vrrp will respond if the master failed just before receiving the packet sent to it. Also master could also fail after receiving the packet but before responding to it i.e during packet processing phase.
 
Any help would be appreciated.
 
Thanks
jahanzeb
_______________________________________________
vrrp mailing list
vrrp <at> ietf.org
https://www1.ietf.org/mailman/listinfo/vrrp
William StanisLaus | 10 Jun 10:25 2004
Picon

RE: failure model

Hello,
    I'm not sure about failure model defined for router failures in vrrp... VRRP itself a election protocol which should overcome shortcomings of router failure.
 
>I want to ask how vrrp will respond if the master failed just before receiving the packet sent to it
 Immedaitely the backup router will become MASTER once it sense the Master has failed, but there is a delay.. if packet hits within that time packet is not serviced, Sender has to retransmit the packet to the new backup router which has become master.
Otherwise, the Backup router will become master to service the packet.
 
>Also master could also fail after receiving the packet but before responding to it i.e during packet processing phase.
In this case the Sender MUST retransmit the packet to the new MASTER router.
 
 
But there are cases/implementations, MASTER and BACKUP routers to kept in sync with the all state machines. In such cases, the packet drop is saved. since both the routers actually accept the packet, but only the MASTER process or takes any action. During failover, the BACKUP becomes MASTER and service the packet.
 
-William.
 
-----Original Message-----
From: vrrp-bounces <at> ietf.org [mailto:vrrp-bounces <at> ietf.org] On Behalf Of Jahanzeb Faizan
Sent: Wednesday, June 09, 2004 7:28 PM
To: vrrp <at> ietf.org
Subject: [vrrp] failure model

 
Hi,
 
I want to know if there is any failure model defined for router failures in vrrp. Also I want to ask how vrrp will respond if the master failed just before receiving the packet sent to it. Also master could also fail after receiving the packet but before responding to it i.e during packet processing phase.
 
Any help would be appreciated.
 
Thanks
jahanzeb
_______________________________________________
vrrp mailing list
vrrp <at> ietf.org
https://www1.ietf.org/mailman/listinfo/vrrp
Radia Perlman | 12 Jun 00:26 2004
Picon

Re: failure model

I believe all these cases wind up the same. Once a sender sends a 
packet, whether the receiver
dies before receiving it, or afterwards but before finishing processing 
it, the packet is lost.
Networking protocols have to be able to cope with lost packets that 
might occur for lots
of reasons. And there are some cases in which packets might get lost for 
a fairly significant
amount of time (where "fairly significant" is not a crisp statement). If 
it takes
some number of seconds for the backup vrrp router to notice and take 
over, then all data packets
sent using the master router will get lost. Again, networks have all 
sorts of issues (like the routing
protocol also might take awhile to notice a failure and to converge to 
correct routes after failures)
where there might be temporary black holes. So applications that run on 
networks have
to be written to be aware of this and be able to cope.

Radia

Jahanzeb Faizan wrote:

>  
> Hi,
>  
> I want to know if there is any failure model defined for router 
> failures in vrrp. Also I want to ask how vrrp will respond if the 
> master failed just before receiving the packet sent to it. Also master 
> could also fail after receiving the packet but before responding to it 
> i.e during packet processing phase.
>  
> Any help would be appreciated.
>  
> Thanks
> jahanzeb

_______________________________________________
vrrp mailing list
vrrp <at> ietf.org
https://www1.ietf.org/mailman/listinfo/vrrp

Internet-Drafts | 17 Jun 17:24 2004
Picon

I-D ACTION:draft-ietf-vrrp-unified-mib-01.txt

A New Internet-Draft is available from the on-line Internet-Drafts directories.
This draft is a work item of the Virtual Router Redundancy Protocol Working Group of the IETF.

	Title		: Definitions of Managed Objects for the VRRPv2 and VRRpv3
	Author(s)	: K. Tata
	Filename	: draft-ietf-vrrp-unified-mib-01.txt
	Pages		: 43
	Date		: 2004-6-17
	
This specification defines a Management Information Base (MIB) for
   use with SNMP-based network management. In particular, it defines
   objects for configuring, monitoring, and controlling routers that
   employ the Virtual Router Redundancy Protocol for both IPv4 and IPv6
   as defined in RFC 3768 and RFC xxxx ( RFC-editor, this is currently
   draft-ietf-vrrp-ipv6-spec-06.txt ). This memo obsoletes RFC 2787.

A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-ietf-vrrp-unified-mib-01.txt

To remove yourself from the I-D Announcement list, send a message to 
i-d-announce-request <at> ietf.org with the word unsubscribe in the body of the message.  
You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce 
to change your subscription settings.

Internet-Drafts are also available by anonymous FTP. Login with the username
"anonymous" and a password of your e-mail address. After logging in,
type "cd internet-drafts" and then
	"get draft-ietf-vrrp-unified-mib-01.txt".

A list of Internet-Drafts directories can be found in
http://www.ietf.org/shadow.html 
or ftp://ftp.ietf.org/ietf/1shadow-sites.txt

Internet-Drafts can also be obtained by e-mail.

Send a message to:
	mailserv <at> ietf.org.
In the body type:
	"FILE /internet-drafts/draft-ietf-vrrp-unified-mib-01.txt".
	
NOTE:	The mail server at ietf.org can return the document in
	MIME-encoded form by using the "mpack" utility.  To use this
	feature, insert the command "ENCODING mime" before the "FILE"
	command.  To decode the response(s), you will need "munpack" or
	a MIME-compliant mail reader.  Different MIME-compliant mail readers
	exhibit different behavior, especially when dealing with
	"multipart" MIME messages (i.e. documents which have been split
	up into multiple messages), so check your local documentation on
	how to manipulate these messages.
		
		
Below is the data which will enable a MIME compliant mail reader
implementation to automatically retrieve the ASCII version of the
Internet-Draft.
Attachment: message/external-body, 140 bytes
Attachment (draft-ietf-vrrp-unified-mib-01.txt): message/external-body, 69 bytes
_______________________________________________
vrrp mailing list
vrrp <at> ietf.org
https://www1.ietf.org/mailman/listinfo/vrrp
jamal | 20 Jun 18:52 2004
Picon
Picon

[Fwd: VRRP and Spanning Tree]

-----Forwarded Message-----

From: Jamal Hadi Salim <hadi <at> znyx.com>
To: vrrp list <vrrp <at> ietf.org>
Subject: VRRP and Spanning Tree
Date: 20 Jun 2004 12:36:44 -0400

We are running into some issues which appear to be synchronization
related when running STP and VRRP in a topology setup we have.
Does anyone have any experience running these two together?
What timers should i be worrying about (eg VRRP heartbeats etc)?
Is this a sane idea to begin with?

cheers,
jamal

_______________________________________________
vrrp mailing list
vrrp <at> ietf.org
https://www1.ietf.org/mailman/listinfo/vrrp

William StanisLaus | 21 Jun 12:33 2004
Picon

RE: [Fwd: VRRP and Spanning Tree]

Yeah there is a issue related to STP and VRRP which we have noticed in
our lab.. If we run VRRP in cluster which is connected using SWITCH
which run's STP, the VRRP failover time is summed with the STP link
failer detection time.
In that case, when we stop STP, VRRP was wrking fine.

-William.

> -----Original Message-----
> From: vrrp-bounces <at> ietf.org [mailto:vrrp-bounces <at> ietf.org] On 
> Behalf Of jamal
> Sent: Sunday, June 20, 2004 10:23 PM
> To: vrrp <at> ietf.org
> Subject: [VRRP] [Fwd: VRRP and Spanning Tree]
> 
> 
> -----Forwarded Message-----
> 
> From: Jamal Hadi Salim <hadi <at> znyx.com>
> To: vrrp list <vrrp <at> ietf.org>
> Subject: VRRP and Spanning Tree
> Date: 20 Jun 2004 12:36:44 -0400
> 
> 
> We are running into some issues which appear to be synchronization
> related when running STP and VRRP in a topology setup we have.
> Does anyone have any experience running these two together?
> What timers should i be worrying about (eg VRRP heartbeats etc)?
> Is this a sane idea to begin with?
> 
> cheers,
> jamal
>  
> 
> 
> 
> 
> _______________________________________________
> vrrp mailing list
> vrrp <at> ietf.org
> https://www1.ietf.org/mailman/listinfo/vrrp
> 

_______________________________________________
vrrp mailing list
vrrp <at> ietf.org
https://www1.ietf.org/mailman/listinfo/vrrp

John.Cruz | 28 Jun 22:55 2004
Picon

VRRP version 3 and IPv6 Neighbor discovery Router Advertisement

Consider the following scenario

------------------------ N
     |        |
     |        |
    R1        R2

If the routers R1 and R2 are not running VRRP,
then both routers send out ND Router advertisements.
With auto-configuration, all hosts on the network N
would choose one of the routers as their default
router. When the current default router fails, the
hosts pick the other one as their default router.

With VRRPv3, only the master sends out ND router
advertisements. All the hosts on the network will set
their default router to the virtual IP address. When
the current master fails, the backup will become the
new master, continue to send out ND router advertisements
and will back-up the virtual IP address currently being
used by the hosts.

All the above makes sense when there is only one virtual
router configured. The VRRPv3 spec suggests that for
load balancing purposes, one can configure multiple VRs.
Let us say that we have two virtual routers configured,
VR1 and VR2, and let the virtual IP addresses for the
virtual router be VIP1 and VIP2. If VIP1 is the interface
address of R1 and VIP2 is the interface address of R2,
the, by default, R1 will be master for VR1 and R2 will
be master for VR2.

Since both R1 and R2 are masters, both will send out
ND router advertisements. All the hosts on the network
will choose one of the advertisements and therefore 
choose one of the routers to be their default router.
If this happens, then there is no load balancing. The
purpose of configuring 2 virtual routers is lost.

If the hosts on the network are not auto-configured, then
we can configure default routes on some hosts to point 
to R1 and the remaining to R2 to acheive load balancing.

I think that coupling ND router advertisements with
VRRP makes sense only if there is one virtual router. If
there are multiple virtual routers, then ND router
advertisements should not be send out as this will mess
with VRRP - basically, hinder load balancing. Note that
in IPv4 case (VRRP version 2), VRRP and Router Discovery
are not coupled.

Any comments?

John

_______________________________________________
vrrp mailing list
vrrp <at> ietf.org
https://www1.ietf.org/mailman/listinfo/vrrp

William StanisLaus | 29 Jun 10:11 2004
Picon

RE: VRRP version 3 and IPv6 Neighbor discovery RouterAdvertisement

Hi,
	I agree with your comments, but there is a race in IPv6 router
discovery, when a host comes up and sents the router solicitation
message for which any router Reponses with the router ip address, in our
case vrrp virtual ip address. Also learnt router ip has a expiration
timer, Once expired there is again a race condition which virtual ip is
configured to the host.
	I agree the worst case is all the host in the subnet is
configured to a single virtual ip address, even though we support load
balance. But we should calculate the probability of such condition.

any thoughts ??

-William.

> -----Original Message-----
> From: vrrp-bounces <at> ietf.org [mailto:vrrp-bounces <at> ietf.org] On 
> Behalf Of John.Cruz <at> nokia.com
> Sent: Tuesday, June 29, 2004 2:26 AM
> To: vrrp <at> ietf.org
> Subject: [VRRP] VRRP version 3 and IPv6 Neighbor discovery 
> RouterAdvertisement
> 
> 
> Consider the following scenario
> 
> ------------------------ N
>      |        |
>      |        |
>     R1        R2
> 
> If the routers R1 and R2 are not running VRRP,
> then both routers send out ND Router advertisements.
> With auto-configuration, all hosts on the network N
> would choose one of the routers as their default
> router. When the current default router fails, the
> hosts pick the other one as their default router.
> 
> With VRRPv3, only the master sends out ND router
> advertisements. All the hosts on the network will set
> their default router to the virtual IP address. When
> the current master fails, the backup will become the
> new master, continue to send out ND router advertisements
> and will back-up the virtual IP address currently being
> used by the hosts.
> 
> All the above makes sense when there is only one virtual
> router configured. The VRRPv3 spec suggests that for
> load balancing purposes, one can configure multiple VRs.
> Let us say that we have two virtual routers configured,
> VR1 and VR2, and let the virtual IP addresses for the
> virtual router be VIP1 and VIP2. If VIP1 is the interface
> address of R1 and VIP2 is the interface address of R2,
> the, by default, R1 will be master for VR1 and R2 will
> be master for VR2.
> 
> Since both R1 and R2 are masters, both will send out
> ND router advertisements. All the hosts on the network
> will choose one of the advertisements and therefore 
> choose one of the routers to be their default router.
> If this happens, then there is no load balancing. The
> purpose of configuring 2 virtual routers is lost.
> 
> If the hosts on the network are not auto-configured, then
> we can configure default routes on some hosts to point 
> to R1 and the remaining to R2 to acheive load balancing.
> 
> I think that coupling ND router advertisements with
> VRRP makes sense only if there is one virtual router. If
> there are multiple virtual routers, then ND router
> advertisements should not be send out as this will mess
> with VRRP - basically, hinder load balancing. Note that
> in IPv4 case (VRRP version 2), VRRP and Router Discovery
> are not coupled.
> 
> Any comments?
> 
> John
> 
> _______________________________________________
> vrrp mailing list
> vrrp <at> ietf.org
> https://www1.ietf.org/mailman/listinfo/vrrp
> 

_______________________________________________
vrrp mailing list
vrrp <at> ietf.org
https://www1.ietf.org/mailman/listinfo/vrrp

John.Cruz | 29 Jun 23:48 2004
Picon

RE: VRRP version 3 and IPv6 Neighbor discovery RouterAdvertisement

I agree that only hosts that are not statically
configured will configure themseleves based on the
ND Router Advertisements. I think that if multiple
VRs are configured for the purpose of load balancing,
then ALL the hosts will be statically configured.
If this is true, then why couple VRRP and ND RA.

My main concern with coupling VRRP and ND router
advertisement in the increased complexity of the ND
implementation. If the situation was "either send
ND RA or don't send", then things are simple. However,
in the presence of multiple VRs, things get complicated.
A router must look at the VRs for which it the master.
If the virtual IP address is something that it is
backing up (because the current master is down), the it must
send out a ND RA on the address owners behalf. It must do this
for each VR. This implies that all routers are configured
similarly. As per the ND RFC, a router will send out only
one RA per interface. In this situation a router may have
to send out more than one RA per interface. I think the
complexity of the ND protocol will be increased just by
coupling it with VRRP.

John

> -----Original Message-----
> From: ext Don Provan [mailto:dprovan <at> bivio.net]
> Sent: Tuesday, June 29, 2004 11:39 AM
> To: Cruz John (Nokia-ES/MTview)
> Cc: vrrp <at> ietf.org
> Subject: RE: [VRRP] VRRP version 3 and IPv6 Neighbor discovery
> RouterAdvertisement
> 
> 
> NDs are sent to pick up any nodes which are not configured
> for load balancing. Nodes configured to use a specific router
> in order to achieve load balancing shouldn't be affected
> by NDs for other routers except when their configured
> router is not available.
> 
> This observation has nothing to do with VRRP: it is just
> as true for a non-VRRP network with redundant routers. So
> I'm not seeing what specifically about VRRP is concerning
> you.
> 
> -don provan
> dprovan <at> bivio.net
> 
> > -----Original Message-----
> > From: vrrp-bounces <at> ietf.org 
> [mailto:vrrp-bounces <at> ietf.org]On Behalf Of
> > John.Cruz <at> nokia.com
> > Sent: Monday, June 28, 2004 1:56 PM
> > To: vrrp <at> ietf.org
> > Subject: [VRRP] VRRP version 3 and IPv6 Neighbor discovery
> > RouterAdvertisement
> > 
> > 
> > Consider the following scenario
> > 
> > ------------------------ N
> >      |        |
> >      |        |
> >     R1        R2
> > 
> > If the routers R1 and R2 are not running VRRP,
> > then both routers send out ND Router advertisements.
> > With auto-configuration, all hosts on the network N
> > would choose one of the routers as their default
> > router. When the current default router fails, the
> > hosts pick the other one as their default router.
> > 
> > With VRRPv3, only the master sends out ND router
> > advertisements. All the hosts on the network will set
> > their default router to the virtual IP address. When
> > the current master fails, the backup will become the
> > new master, continue to send out ND router advertisements
> > and will back-up the virtual IP address currently being
> > used by the hosts.
> > 
> > All the above makes sense when there is only one virtual
> > router configured. The VRRPv3 spec suggests that for
> > load balancing purposes, one can configure multiple VRs.
> > Let us say that we have two virtual routers configured,
> > VR1 and VR2, and let the virtual IP addresses for the
> > virtual router be VIP1 and VIP2. If VIP1 is the interface
> > address of R1 and VIP2 is the interface address of R2,
> > the, by default, R1 will be master for VR1 and R2 will
> > be master for VR2.
> > 
> > Since both R1 and R2 are masters, both will send out
> > ND router advertisements. All the hosts on the network
> > will choose one of the advertisements and therefore 
> > choose one of the routers to be their default router.
> > If this happens, then there is no load balancing. The
> > purpose of configuring 2 virtual routers is lost.
> > 
> > If the hosts on the network are not auto-configured, then
> > we can configure default routes on some hosts to point 
> > to R1 and the remaining to R2 to acheive load balancing.
> > 
> > I think that coupling ND router advertisements with
> > VRRP makes sense only if there is one virtual router. If
> > there are multiple virtual routers, then ND router
> > advertisements should not be send out as this will mess
> > with VRRP - basically, hinder load balancing. Note that
> > in IPv4 case (VRRP version 2), VRRP and Router Discovery
> > are not coupled.
> > 
> > Any comments?
> > 
> > John
> > 
> > _______________________________________________
> > vrrp mailing list
> > vrrp <at> ietf.org
> > https://www1.ietf.org/mailman/listinfo/vrrp
> 

_______________________________________________
vrrp mailing list
vrrp <at> ietf.org
https://www1.ietf.org/mailman/listinfo/vrrp


Gmane