Mailing List Archive: Redundant REs: annoying features et al.

Redundant REs: annoying features et al.

Oct 28, 2003, 4:17 AM

Post #1 of 8 (1684 views)

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello, guys. We have a lot of Junipers deployed, every one of them with two
REs.

I found that the backup routing engine is all the time trying to connect to
NTP servers, trying to upload its configuration, and so on. Even more, when
you switch routing engines, the backup one is out of date, due to the fact
that was unable to sync with the NTP server.

Is there any way of solving this issue? Maybe the new statement [ system
backup-router ] will be useful? But I cannot manage to understand it in order
to set it up for fxp1/2 (internal card, whatever name it has).

Is it possible?

Regards,

Ender.
- --
You can't even be a vegetable, because even a artichoke has heart.
-- Am?lie (Le fabuleux destin d'Am?lie Poulain)
- --
Servicios de red - Network services
Centro de Comunicaciones CSIC/RedIRIS
Spanish Academic Network for Research and Development
Madrid (Spain)
Tlf (+34) 91.585.49.05
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)

iD8DBQE/njOaWs/EhA1iABsRAvJqAJ9RXmsdLk+uYRtyp00tkBN0h0BhYQCgyTRB
YLsSXucDmNoxMKvZV0DxZeM=
=142S
-----END PGP SIGNATURE-----

Redundant REs: annoying features et al. [ In reply to ]

dek at hades

Oct 28, 2003, 6:56 PM

Post #2 of 8 (1668 views)

Permalink

Hi David,

groups {
re0 {
system {
host-name router_re0;
}
interfaces {
fxp0 {
unit 0 {
family inet {
address 10.0.0.1/24 {
}
}
}
}
}
re1 {
system {
host-name router_re1;
}
interfaces {
fxp0 {
unit 0 {
family inet {
address 10.0.0.2/24 {
}
}
}
}
}
}

This config, when applied to re0 and re1, will configure their fxp0 with
different IP addresses, so they both can talk to the network. Just note that
on an inactive re the routing process isn't active.

Hope this is what you're after.

SY,

-- D.K.

On Tue, Oct 28, 2003 at 11:15:06AM +0200, David Martinez CSIC RedIRIS wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Hello, guys. We have a lot of Junipers deployed, every one of them with two
> REs.
>
> I found that the backup routing engine is all the time trying to connect to
> NTP servers, trying to upload its configuration, and so on. Even more, when
> you switch routing engines, the backup one is out of date, due to the fact
> that was unable to sync with the NTP server.
>
> Is there any way of solving this issue? Maybe the new statement [ system
> backup-router ] will be useful? But I cannot manage to understand it in order
> to set it up for fxp1/2 (internal card, whatever name it has).
>
> Is it possible?
>
> Regards,
>
>
> Ender.
> - --
> You can't even be a vegetable, because even a artichoke has heart.
> -- Am?lie (Le fabuleux destin d'Am?lie Poulain)
> - --
> Servicios de red - Network services
> Centro de Comunicaciones CSIC/RedIRIS
> Spanish Academic Network for Research and Development
> Madrid (Spain)
> Tlf (+34) 91.585.49.05
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.2.3 (GNU/Linux)
>
> iD8DBQE/njOaWs/EhA1iABsRAvJqAJ9RXmsdLk+uYRtyp00tkBN0h0BhYQCgyTRB
> YLsSXucDmNoxMKvZV0DxZeM=
> =142S
> -----END PGP SIGNATURE-----
>
>
> _______________________________________________
> juniper-nsp mailing list juniper-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/juniper-nsp
---end quoted text---

--
D.K.

Redundant REs: annoying features et al. [ In reply to ]

gtate at juniper

Oct 28, 2003, 11:38 PM

Post #3 of 8 (1663 views)

Permalink

This is the correct configuration. It is also well to use the backup
router statement

set system backup-router <default-gateway-for-fxp0-network>

Bellow lets assume:
set system backup-router 10.0.0.254

This will allow you to communicate to the secondary RE via the fxp0 if
you wish without logging in from other RE

Gary

On Tuesday, Oct 28, 2003, at 15:56 US/Pacific, Dmitri Kalintsev wrote:

> Hi David,
>
> groups {
> re0 {
> system {
> host-name router_re0;
> }
> interfaces {
> fxp0 {
> unit 0 {
> family inet {
> address 10.0.0.1/24 {
> }
> }
> }
> }
> }
> re1 {
> system {
> host-name router_re1;
> }
> interfaces {
> fxp0 {
> unit 0 {
> family inet {
> address 10.0.0.2/24 {
> }
> }
> }
> }
> }
> }
>
> This config, when applied to re0 and re1, will configure their fxp0
> with
> different IP addresses, so they both can talk to the network. Just
> note that
> on an inactive re the routing process isn't active.
>
> Hope this is what you're after.
>
> SY,
>
> -- D.K.
>
> On Tue, Oct 28, 2003 at 11:15:06AM +0200, David Martinez CSIC RedIRIS
> wrote:
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>> Hello, guys. We have a lot of Junipers deployed, every one of them
>> with two
>> REs.
>>
>> I found that the backup routing engine is all the time trying to
>> connect to
>> NTP servers, trying to upload its configuration, and so on. Even
>> more, when
>> you switch routing engines, the backup one is out of date, due to the
>> fact
>> that was unable to sync with the NTP server.
>>
>> Is there any way of solving this issue? Maybe the new statement [
>> system
>> backup-router ] will be useful? But I cannot manage to understand it
>> in order
>> to set it up for fxp1/2 (internal card, whatever name it has).
>>
>> Is it possible?
>>
>> Regards,
>>
>>
>> Ender.
>> - --
>> You can't even be a vegetable, because even a artichoke has heart.
>> -- Am?lie (Le fabuleux destin d'Am?lie Poulain)
>> - --
>> Servicios de red - Network services
>> Centro de Comunicaciones CSIC/RedIRIS
>> Spanish Academic Network for Research and Development
>> Madrid (Spain)
>> Tlf (+34) 91.585.49.05
>> -----BEGIN PGP SIGNATURE-----
>> Version: GnuPG v1.2.3 (GNU/Linux)
>>
>> iD8DBQE/njOaWs/EhA1iABsRAvJqAJ9RXmsdLk+uYRtyp00tkBN0h0BhYQCgyTRB
>> YLsSXucDmNoxMKvZV0DxZeM=
>> =142S
>> -----END PGP SIGNATURE-----
>>
>>
>> _______________________________________________
>> juniper-nsp mailing list juniper-nsp@puck.nether.net
>> http://puck.nether.net/mailman/listinfo/juniper-nsp
> ---end quoted text---
>
> --
> D.K.
> _______________________________________________
> juniper-nsp mailing list juniper-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/juniper-nsp
>

Redundant REs: annoying features et al. [ In reply to ]

gtate at juniper

Oct 29, 2003, 1:04 AM

Post #4 of 8 (1665 views)

Permalink

Let me know if you need the NTP services.

By the way I am assuming you are using commit sync? If you still have
problems you can send an abridged system config to my personal mail.

Gary

On Tuesday, Oct 28, 2003, at 20:38 US/Pacific, Gary Tate wrote:

> This is the correct configuration. It is also well to use the backup
> router statement
>
> set system backup-router <default-gateway-for-fxp0-network>
>
> Bellow lets assume:
> set system backup-router 10.0.0.254
>
> This will allow you to communicate to the secondary RE via the fxp0 if
> you wish without logging in from other RE
>
> Gary
>
> On Tuesday, Oct 28, 2003, at 15:56 US/Pacific, Dmitri Kalintsev wrote:
>
>> Hi David,
>>
>> groups {
>> re0 {
>> system {
>> host-name router_re0;
>> }
>> interfaces {
>> fxp0 {
>> unit 0 {
>> family inet {
>> address 10.0.0.1/24 {
>> }
>> }
>> }
>> }
>> }
>> re1 {
>> system {
>> host-name router_re1;
>> }
>> interfaces {
>> fxp0 {
>> unit 0 {
>> family inet {
>> address 10.0.0.2/24 {
>> }
>> }
>> }
>> }
>> }
>> }
>>
>> This config, when applied to re0 and re1, will configure their fxp0
>> with
>> different IP addresses, so they both can talk to the network. Just
>> note that
>> on an inactive re the routing process isn't active.
>>
>> Hope this is what you're after.
>>
>> SY,
>>
>> -- D.K.
>>
>> On Tue, Oct 28, 2003 at 11:15:06AM +0200, David Martinez CSIC RedIRIS
>> wrote:
>>> -----BEGIN PGP SIGNED MESSAGE-----
>>> Hash: SHA1
>>>
>>> Hello, guys. We have a lot of Junipers deployed, every one of them
>>> with two
>>> REs.
>>>
>>> I found that the backup routing engine is all the time trying to
>>> connect to
>>> NTP servers, trying to upload its configuration, and so on. Even
>>> more, when
>>> you switch routing engines, the backup one is out of date, due to
>>> the fact
>>> that was unable to sync with the NTP server.
>>>
>>> Is there any way of solving this issue? Maybe the new statement [
>>> system
>>> backup-router ] will be useful? But I cannot manage to understand it
>>> in order
>>> to set it up for fxp1/2 (internal card, whatever name it has).
>>>
>>> Is it possible?
>>>
>>> Regards,
>>>
>>>
>>> Ender.
>>> - --
>>> You can't even be a vegetable, because even a artichoke has heart.
>>> -- Am?lie (Le fabuleux destin d'Am?lie Poulain)
>>> - --
>>> Servicios de red - Network services
>>> Centro de Comunicaciones CSIC/RedIRIS
>>> Spanish Academic Network for Research and Development
>>> Madrid (Spain)
>>> Tlf (+34) 91.585.49.05
>>> -----BEGIN PGP SIGNATURE-----
>>> Version: GnuPG v1.2.3 (GNU/Linux)
>>>
>>> iD8DBQE/njOaWs/EhA1iABsRAvJqAJ9RXmsdLk+uYRtyp00tkBN0h0BhYQCgyTRB
>>> YLsSXucDmNoxMKvZV0DxZeM=
>>> =142S
>>> -----END PGP SIGNATURE-----
>>>
>>>
>>> _______________________________________________
>>> juniper-nsp mailing list juniper-nsp@puck.nether.net
>>> http://puck.nether.net/mailman/listinfo/juniper-nsp
>> ---end quoted text---
>>
>> --
>> D.K.
>> _______________________________________________
>> juniper-nsp mailing list juniper-nsp@puck.nether.net
>> http://puck.nether.net/mailman/listinfo/juniper-nsp
>>
>

Redundant REs: annoying features et al. [ In reply to ]

david.martinez at rediris

Oct 30, 2003, 6:04 AM

Post #5 of 8 (1676 views)

Permalink

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

El Mi?rcoles, 29 de Octubre de 2003 07:03, Gary Tate escribi?:
> Let me know if you need the NTP services.
>
> By the way I am assuming you are using commit sync? If you still have
> problems you can send an abridged system config to my personal mail.

The fact is that I always supposed that ntpd (or other services) in the
backup RE could use the ntpd in the working RE through fxp2. Is that trick
the system backup-router? Is necessary to have a hub for example connecting
the two RE fxp0 for this solution?

Regards,

Ender.
- --
And need I remind you that I am naked in the snow...? I
can't feel any of my extremities, and I mean *any* of them...
-- Skinner (The League of Extraordinary Gentlemen)
- --
Servicios de red - Network services
Centro de Comunicaciones CSIC/RedIRIS
Spanish Academic Network for Research and Development
Madrid (Spain)
Tlf (+34) 91.585.49.05
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)

iD8DBQE/oPA4Ws/EhA1iABsRApmEAJkBRoIqik1o2oDEzun2FNfkxI2TogCgsiiN
VaOrpMGmdxSmcuRdjOz5LXU=
=1cL0
-----END PGP SIGNATURE-----

Redundant REs: annoying features et al. [ In reply to ]

harry at juniper

Oct 30, 2003, 10:23 AM

Post #6 of 8 (1662 views)

Permalink

The backup-router has nothing to do with RE synchronization as I understand.
The REs can talk internally w/o an fxp0 based OoB network. The backup router
statement identifies a default gateway for OoB routing that is used *only*
when rpd is not running, such as when the router has just booted or in the
case of rpd thrash.

> -----Original Message-----
> From: juniper-nsp-bounces@puck.nether.net
> [mailto:juniper-nsp-bounces@puck.nether.net] On Behalf Of
> David Martinez CSIC RedIRIS
> Sent: Thursday, October 30, 2003 3:04 AM
> To: Gary Tate
> Cc: juniper-nsp@puck.nether.net; Dmitri Kalintsev
> Subject: Re: [j-nsp] Redundant REs: annoying features et al.
>
>
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> El Mi?rcoles, 29 de Octubre de 2003 07:03, Gary Tate escribi?:
> > Let me know if you need the NTP services.
> >
> > By the way I am assuming you are using commit sync? If you
> still have
> > problems you can send an abridged system config to my personal mail.
>
> The fact is that I always supposed that ntpd (or other
> services) in the
> backup RE could use the ntpd in the working RE through fxp2.
> Is that trick
> the system backup-router? Is necessary to have a hub for
> example connecting
> the two RE fxp0 for this solution?
>
> Regards,
>
>
> Ender.
> - --
> And need I remind you that I am naked in the snow...? I
> can't feel any of my extremities, and I mean *any* of them...
> -- Skinner (The League of Extraordinary Gentlemen)
> - --
> Servicios de red - Network services
> Centro de Comunicaciones CSIC/RedIRIS
> Spanish Academic Network for Research and Development
> Madrid (Spain)
> Tlf (+34) 91.585.49.05
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.2.3 (GNU/Linux)
>
> iD8DBQE/oPA4Ws/EhA1iABsRApmEAJkBRoIqik1o2oDEzun2FNfkxI2TogCgsiiN
> VaOrpMGmdxSmcuRdjOz5LXU=
> =1cL0
> -----END PGP SIGNATURE-----
>
>
> _______________________________________________
> juniper-nsp mailing list juniper-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/junipe> r-nsp
>

Redundant REs: annoying features et al. [ In reply to ]

gtate at juniper

Nov 3, 2003, 3:28 PM

Post #7 of 8 (1664 views)

Permalink

NTP is not synchronized across the internal Ethernet. It is synced as
per the ntp configuration, which, I assume, is configure the same on
each RE in this scenario because you are using commit sync.

If you have the backup-router configured then the backup re will source
packets from the fxp0 address to the backup-router (of course only if
you have the fxp0 connected). If the connection and configuration is
correct then the NTP sync and associations should be up on the backup
re.

Gary

On Thursday, Oct 30, 2003, at 03:04 US/Pacific, David Martinez CSIC
RedIRIS wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> El Mi?rcoles, 29 de Octubre de 2003 07:03, Gary Tate escribi?:
>> Let me know if you need the NTP services.
>>
>> By the way I am assuming you are using commit sync? If you still have
>> problems you can send an abridged system config to my personal mail.
>
> The fact is that I always supposed that ntpd (or other services) in
> the
> backup RE could use the ntpd in the working RE through fxp2. Is that
> trick
> the system backup-router? Is necessary to have a hub for example
> connecting
> the two RE fxp0 for this solution?
>
> Regards,
>
>
> Ender.
> - --
> And need I remind you that I am naked in the snow...? I
> can't feel any of my extremities, and I mean *any* of them...
> -- Skinner (The League of Extraordinary Gentlemen)
> - --
> Servicios de red - Network services
> Centro de Comunicaciones CSIC/RedIRIS
> Spanish Academic Network for Research and Development
> Madrid (Spain)
> Tlf (+34) 91.585.49.05
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.2.3 (GNU/Linux)
>
> iD8DBQE/oPA4Ws/EhA1iABsRApmEAJkBRoIqik1o2oDEzun2FNfkxI2TogCgsiiN
> VaOrpMGmdxSmcuRdjOz5LXU=
> =1cL0
> -----END PGP SIGNATURE-----
>

Redundant REs: annoying features et al. [ In reply to ]

david.martinez at rediris

Nov 19, 2003, 2:43 AM

Post #8 of 8 (1666 views)

Permalink

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

El Lunes, 3 de Noviembre de 2003 21:28, Gary Tate escribi?:
> NTP is not synchronized across the internal Ethernet. It is synced as
> per the ntp configuration, which, I assume, is configure the same on
> each RE in this scenario because you are using commit sync.
>
> If you have the backup-router configured then the backup re will source
> packets from the fxp0 address to the backup-router (of course only if
> you have the fxp0 connected). If the connection and configuration is
> correct then the NTP sync and associations should be up on the backup
> re.

Sorry for not letting you know my greetings. Thank you all for your replies.
I've set backup-router statements in the affected routers and they're working
properly now. It was only a misunderstanding problem.

Gary, the doc you sent me was good, thank you.

Regards,

Ender.
- --
Mr. Anderson! Welcome back, we missed you.
-- Agent Smith (Matrix Revolutions)
- --
Servicios de red - Network services
Centro de Comunicaciones CSIC/RedIRIS
Spanish Academic Network for Research and Development
Madrid (Spain)
Tlf (+34) 91.585.49.05
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)

iD8DBQE/ux8HWs/EhA1iABsRAo24AJsEXekYSwGpVhs5nbZU2NOLEQ1UawCgzBGG
Ok7JOH9WFyhWM2TMqNoW3Eg=
=HuHO
-----END PGP SIGNATURE-----