Mailing List Archive: Quagga: Users

(no subject)

Index | Next | Previous | View Threaded

philipp-paeper at rzserv1-sud

Oct 13, 2003, 1:53 PM

Post #1 of 12 (2227 views)
Permalink

(no subject)

hello,

i have to set up a small test-network with 3 routers. they should use ospf.

sometimes the routers have the right routes (after restarting the network)
(linux cli --> route), but after a minute ore 30 sec they dissapear...

you can find the network plan,config-files and some output under: http://paeper.
luenecom.de/routing/

gw3 should have gw1 (10.1.1.1) or gw2 (10.1.3.2) as router for 10.1.2.0 ...

i hope you can help me

best regards
piepre

quagga at gushi

Oct 22, 2003, 7:54 AM

Post #2 of 12 (2139 views)
Permalink

(no subject) [In reply to]

How can I make it so that I can log directly into a unified interface from
the outside? Is there any way to make the zebra login on port 2601 behave
like vtysh? I essentially would like to give my cisco people the best
representation of a cisco-like interface I can.

-Dan Mahoney

--

"Your future hasn't been written yet; no one's has. So make it a good
one!"

-"Doc" Emmet L. Browne, Back to the Future III

--------Dan Mahoney--------
Techie, Sysadmin, WebGeek
Gushi on efnet/undernet IRC
ICQ: 13735144 AIM: LarpGM
Site: http://www.gushi.org
---------------------------

teun at moonblade

Oct 22, 2003, 8:46 AM

Post #3 of 12 (2120 views)
Permalink

Re: (no subject) [In reply to]

----- Original Message -----
From: "Dan Mahoney, System Admin" <quagga [at] gushi>
To: "Quagga-Users [at] Lists Quagga. Net" <quagga-users [at] lists>
Sent: Wednesday, October 22, 2003 4:54 PM
Subject: [quagga-users 704] (no subject)

> How can I make it so that I can log directly into a unified interface from
> the outside? Is there any way to make the zebra login on port 2601 behave
> like vtysh? I essentially would like to give my cisco people the best
> representation of a cisco-like interface I can.
>
> -Dan Mahoney
>

Never tried it, but you might try to set vtysh as the login shell for one
specific user and use that account. But what's so hard on ssh'ing to a
box and typing 'vtysh' ?

Teun

>> spamtrap: spam [at] moonblade - do NOT ever send email to this address <<

paul at clubi

Oct 22, 2003, 8:50 AM

Post #4 of 12 (2135 views)
Permalink

Re: (no subject) [In reply to]

On Wed, 22 Oct 2003, Dan Mahoney, System Admin wrote:

> How can I make it so that I can log directly into a unified
> interface from the outside? Is there any way to make the zebra
> login on port 2601 behave like vtysh?

create a vtysh user, with whatever you specified to --vtysh-group as
its group and vtysh as its shell.

alternatively, and probably better, run vtysh from inetd as user
vtysh / group = --vtysh-group on some port (should work i think).

You can give individual people individual login passwords via PAM.

> I essentially would like to give my cisco people the best
> representation of a cisco-like interface I can.
>
> -Dan Mahoney

regards,
--
Paul Jakma paul [at] clubi paul [at] jakma Key ID: 64A2FF6A
warning: do not ever send email to spam [at] dishone
Fortune:
Before Xerox, five carbons were the maximum extension of anybody's ego.

ngrundy at isoproplex

Oct 22, 2003, 7:02 PM

Post #5 of 12 (2136 views)
Permalink

Re: (no subject) [In reply to]

I setup the quagga user to have a shell of /usr/lobal/bin/vtysh (path varies of course)
and just ssh into the machine with the credentials of that user.

> How can I make it so that I can log directly into a unified interface from
> the outside? Is there any way to make the zebra login on port 2601 behave
> like vtysh? I essentially would like to give my cisco people the best
> representation of a cisco-like interface I can.
>
> -Dan Mahoney
>
>
> --
>
> "Your future hasn't been written yet; no one's has. So make it a good
> one!"
>
> -"Doc" Emmet L. Browne, Back to the Future III
>
> --------Dan Mahoney--------
> Techie, Sysadmin, WebGeek
> Gushi on efnet/undernet IRC
> ICQ: 13735144 AIM: LarpGM
> Site: http://www.gushi.org
> ---------------------------
>
>
> _______________________________________________
> Quagga-users mailing list
> Quagga-users [at] lists
> http://lists.quagga.net/mailman/listinfo/quagga-users
>

~Nick
-----
Every cloud has a silver lining (except for the mushroom shaped ones, which have a lining of Iridium & Strontium 90)

paul at clubi

Jan 9, 2006, 3:40 PM

Post #6 of 12 (2141 views)
Permalink

Re: (no subject) [In reply to]

On Mon, 9 Jan 2006, Jason Sigurdur wrote:

> going throught site C. If the standard metric for fastetherenet is
> 20 then going from a to c to b would be '40'? would'nt the lower
> cost of 30 from A. to B. on the second tunnel be used?

Yes. Doesn't it?

> Also, if possible any suggestions on setting up area's. Each Site
> has a LAN behind it. For example site A. would have 10.100.0.0 B.
> 10.200.0.0 . Should each network be setup as a stub or should I
> just keep everything on area 0?

The LANs, will you have anything on them that needs to speak OSPF? If
not, no point setting up any areas.

If yes, you need to figure out how whether each 'area' will ever need
to redistribute external routes into OSPF.

Stub areas can help keep your OSPF domain a bit 'cleaner' though.

regards,
--
Paul Jakma paul [at] clubi paul [at] jakma Key ID: 64A2FF6A
Fortune:
The root of all superstition is that men observe when a thing hits,
but not when it misses.
-- Francis Bacon
_______________________________________________
Quagga-users mailing list
Quagga-users [at] lists
http://lists.quagga.net/mailman/listinfo/quagga-users

Jason.Sigurdur at aspenview

Jan 10, 2006, 8:31 AM

Post #7 of 12 (2119 views)
Permalink

Re: (no subject) [In reply to]

Hi, thanks you for the quick reply.

I figured out my problem. A metric of '10' is given to each
Fast-Ethernet interface, therefore the metric from a to b would be 20.

Originally the secondary tunnel interface has a metric of 30 which would
be a equal metric from a(10) -> b(10) -> c(10) '30' . So I changed my
'ip ospf cost' to 25. This seemed to work. Am I correct with this
assumption?

Just one more question. I also noticed that all my p-to-p routes have
'equalize' beside them when doing an 'ip route'; there is no nexthop
because the other route has a higher metric? Is this because there are
two possible routes to the same network? In this case only one being
used at a time.

Thx jason

-----Original Message-----
From: Paul Jakma [mailto:paul [at] clubi]
Sent: Monday, January 09, 2006 4:41 PM
To: Jason Sigurdur
Cc: quagga-users [at] lists
Subject: Re: [quagga-users 6275] (no subject)

On Mon, 9 Jan 2006, Jason Sigurdur wrote:

> going throught site C. If the standard metric for fastetherenet is
> 20 then going from a to c to b would be '40'? would'nt the lower
> cost of 30 from A. to B. on the second tunnel be used?

Yes. Doesn't it?

> Also, if possible any suggestions on setting up area's. Each Site
> has a LAN behind it. For example site A. would have 10.100.0.0 B.
> 10.200.0.0 . Should each network be setup as a stub or should I
> just keep everything on area 0?

The LANs, will you have anything on them that needs to speak OSPF? If
not, no point setting up any areas.

If yes, you need to figure out how whether each 'area' will ever need
to redistribute external routes into OSPF.

Stub areas can help keep your OSPF domain a bit 'cleaner' though.

regards,
--
Paul Jakma paul [at] clubi paul [at] jakma Key ID: 64A2FF6A
Fortune:
The root of all superstition is that men observe when a thing hits,
but not when it misses.
-- Francis Bacon

_______________________________________________
Quagga-users mailing list
Quagga-users [at] lists
http://lists.quagga.net/mailman/listinfo/quagga-users

paul at clubi

Jan 10, 2006, 10:19 AM

Post #8 of 12 (2115 views)
Permalink

Re: (no subject) [In reply to]

On Tue, 10 Jan 2006, Jason Sigurdur wrote:

> Hi, thanks you for the quick reply.
>
> I figured out my problem. A metric of '10' is given to each
> Fast-Ethernet interface, therefore the metric from a to b would be
> 20.

It'd be 10.

> Originally the secondary tunnel interface has a metric of 30 which
> would be a equal metric from a(10) -> b(10) -> c(10) '30' . So I
> changed my 'ip ospf cost' to 25. This seemed to work. Am I correct
> with this assumption?

No, because the cost for C's link doesnt factor into it.

a(10)->b(10)->c(10): a->b + b->c is 10+10 = 20.

The direct a->b link is just 10.

So, if you'd prefer the secondary link to be used over a 'via
neighbour' path, you should set its cost to 15 probably.

> Just one more question. I also noticed that all my p-to-p routes
> have 'equalize' beside them when doing an 'ip route';

Yeah, that's some kind of Linux netlink ECMP flag (it's supposed to
make packets hitting route do per-packet ECMP, rather than per-flow,
but apparently it isn't implemented - it does nothing). We don't set
that flag in 0.99 anymore.

> there is no nexthop because the other route has a higher metric?

I don't understand this question. There should always be a nexthop
(or multiple ones, if an ECMP route). Are you maybe doing 'ip route |
grep <prefix>', ECMP routes will have the nexthops on seperate lines.

> Is this because there are two possible routes to the same network?

If there are two nexthops, then yes, because there are two possible
routes.

> In this case only one being used at a time.

Not completely sure about what you mean here either. If you mean
you've observed packets only going out via one of the nexthops, it's
likely because of flow-caching, each flow will use the same nexthop.
Try setting up more flows.

What exactly constitutes a flow on Linux, i'm not sure.

> Thx jason

regards,
--
Paul Jakma paul [at] clubi paul [at] jakma Key ID: 64A2FF6A
Fortune:
Dying is a very dull, dreary affair. My advice to you is to have
nothing whatever to do with it.
-- W. Somerset Maughm, his last words
_______________________________________________
Quagga-users mailing list
Quagga-users [at] lists
http://lists.quagga.net/mailman/listinfo/quagga-users

Jason.Sigurdur at aspenview

Jan 10, 2006, 11:23 AM

Post #9 of 12 (2135 views)
Permalink

Re: (no subject) [In reply to]

Hi, for the 'ip ospf cost' it was 15 not 25.

Thank you for clearing things up.

Jason

-----Original Message-----
From: Paul Jakma [mailto:paul [at] clubi]
Sent: Tuesday, January 10, 2006 11:19 AM
To: Jason Sigurdur
Cc: quagga-users [at] lists
Subject: Re: [quagga-users 6278] Re: (no subject)

On Tue, 10 Jan 2006, Jason Sigurdur wrote:

> Hi, thanks you for the quick reply.
>
> I figured out my problem. A metric of '10' is given to each
> Fast-Ethernet interface, therefore the metric from a to b would be
> 20.

It'd be 10.

> Originally the secondary tunnel interface has a metric of 30 which
> would be a equal metric from a(10) -> b(10) -> c(10) '30' . So I
> changed my 'ip ospf cost' to 25. This seemed to work. Am I correct
> with this assumption?

No, because the cost for C's link doesnt factor into it.

a(10)->b(10)->c(10): a->b + b->c is 10+10 = 20.

The direct a->b link is just 10.

So, if you'd prefer the secondary link to be used over a 'via
neighbour' path, you should set its cost to 15 probably.

> Just one more question. I also noticed that all my p-to-p routes
> have 'equalize' beside them when doing an 'ip route';

Yeah, that's some kind of Linux netlink ECMP flag (it's supposed to
make packets hitting route do per-packet ECMP, rather than per-flow,
but apparently it isn't implemented - it does nothing). We don't set
that flag in 0.99 anymore.

> there is no nexthop because the other route has a higher metric?

I don't understand this question. There should always be a nexthop
(or multiple ones, if an ECMP route). Are you maybe doing 'ip route |
grep <prefix>', ECMP routes will have the nexthops on seperate lines.

> Is this because there are two possible routes to the same network?

If there are two nexthops, then yes, because there are two possible
routes.

> In this case only one being used at a time.

Not completely sure about what you mean here either. If you mean
you've observed packets only going out via one of the nexthops, it's
likely because of flow-caching, each flow will use the same nexthop.
Try setting up more flows.

What exactly constitutes a flow on Linux, i'm not sure.

> Thx jason

regards,
--
Paul Jakma paul [at] clubi paul [at] jakma Key ID: 64A2FF6A
Fortune:
Dying is a very dull, dreary affair. My advice to you is to have
nothing whatever to do with it.
-- W. Somerset Maughm, his last words

_______________________________________________
Quagga-users mailing list
Quagga-users [at] lists
http://lists.quagga.net/mailman/listinfo/quagga-users

smn at libero

Mar 29, 2007, 8:38 AM

Post #10 of 12 (1878 views)
Permalink

Re: (no subject) [In reply to]

Thanks Andrew for your prompt reply!

I think your information will be very useful.

Simone

---------- Initial Header -----------

>From : "Andrew J. Schorr" aschorr [at] telemetry-investments
To : "smn [at] libero" smn [at] libero
Cc : "quagga-users" quagga-users [at] lists
Date : Thu, 29 Mar 2007 11:06:04 -0400
Subject : Re: [quagga-users 8195] (no subject)

> Hi Simone,
>
> On Thu, Mar 29, 2007 at 04:24:34PM +0200, smn [at] libero wrote:
> > I'm a new Quagga user and I am not sure if to launch also the watchquagga task besides the router deamons.
>
> It is your choice whether to use watchquagga. There are 3 basic
> ways to start the daemons:
>
> 1. Start each daemon individually, and do not run watchquagga.
> 2. Start watchquagga, and let it start all the other daemons.
> 3. Both: start each daemon, and then start watchquagga. Since
> watchquagga will notice that the daemons are already running, it
> will not start them.
>
> What works best for you may depend on your platform and
> its standard techniques for starting daemons. There has
> been some discussion of trying to use watchquagga as the
> canonical way to run the daemons (i.e. push everyone towards
> approach #2 above)...
>
> > I think this can be a useful way to run routing protocol in a safer way, and it seems to me that mode 4 (number is taken from usage function) is the best and safest.
>
> I agree, that is how I use watchquagga (in mode 4). In the future,
> if/when the quagga daemons are enhanced to support more sophisticated
> resyncing capabilities, other modes may be preferable...
>
> > Has this usage mode some known drawbacks?
>
> I don't think so. On my redhat linux box, I have the following
> in /etc/sysconfig/quagga:
>
> WATCH_OPTS="-Az -b_ -r/sbin/service_%s_restart -s/sbin/service_%s_start -k/sbin/service_%s_stop"
> WATCH_DAEMONS="zebra ospfd"
>
> This works fine for me (running only ospfd on my core routers).
> On my ASBR boxes, I have:
>
> WATCH_DAEMONS="zebra ospfd bgpd ripd"
>
> Please let us know if you encounter any problems.
>
> Regards,
> Andy
>

------------------------------------------------------
Leggi GRATIS le tue mail con il telefonino i-mode� di Wind
http://click.libero.it/imode

_______________________________________________
Quagga-users mailing list
Quagga-users [at] lists
http://lists.quagga.net/mailman/listinfo/quagga-users

tyler at tolaris

Nov 3, 2010, 12:54 AM

Post #11 of 12 (1156 views)
Permalink

Re: (no subject) [In reply to]

And under Virtualbox too. Create virtual machines and virtual networks
between them, such that some VMs are only accessible networked through
others, and use Quagga to enable that. Plan it just as you would a
physical network.

Regards,
Tyler

On Tue, 2010-11-02 at 17:41 -0400, Jack Carrozzo wrote:
> It works fine on linux or bsd under VMware or xen. Use the standard
> docs.
>
>
> -Jack Carrozzo
>
> On Tue, Nov 2, 2010 at 4:59 PM, johnny aiyes
> <whoisdatjohnny [at] yahoo> wrote:
> hello quagga team.....please how do i configure quagga in a
> virtual environment to show the performance of OSPF and RIP in
> a multi-router environment such as VMware
>
>
>
> _______________________________________________
> Quagga-users mailing list
> Quagga-users [at] lists
> http://lists.quagga.net/mailman/listinfo/quagga-users
>
>
>
> _______________________________________________
> Quagga-users mailing list
> Quagga-users [at] lists
> http://lists.quagga.net/mailman/listinfo/quagga-users

--
"Offending fundamentalists isn't my goal – but if it is an inevitable
side-effect of defending human rights, so be it."
-- Johann Hari

_______________________________________________
Quagga-users mailing list
Quagga-users [at] lists
http://lists.quagga.net/mailman/listinfo/quagga-users

pitfloner at hotmail

Jul 12, 2012, 4:49 PM

Post #12 of 12 (631 views)
Permalink

Re: (no subject) [In reply to]

. http://heartsnflowersplus.com/atvvvxxx.php?ifortuneid=21xf

Index | Next | Previous | View Threaded

Interested in having your list archived? Contact Gossamer Threads