[j-nsp] QFX5100 vs ACX5048

[Colton Conor]

What is the main difference between these two boxes? Hardware wise they
look identical. Is there anything on the hardware side that makes the
ACX5048 better than a QFX5100, or is it only software related?
_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
https://puck.nether.net/mailman/listinfo/juniper-nsp

[Alexandre Guimaraes]

Better in terms of concept. In term of usage, i still investing in qfx5100

Acx5058 Suppose to be a promise of a new future, unfortunately, with all problematic of the qfx5100 hardware, the acx5048 leak vlan till the last breath of cpu.... after that, all deamons and services going down.... up and down, up and down.

I never more brought one peace ACX5048 after jtac didnt responds why and solution for the leaking...( I have only two acx5048 and hundreds on QFX...).

The new promise is the new acx5448. No vlan leaking, a good load balance(ae) algorithm, full of this.... full of that.... a lot of promise.

Let’s see...

att
Alexandre

Em 1 de jul de 2018, à(s) 19:31, Colton Conor <***@gmail.com> escreveu:

> What is the main difference between these two boxes? Hardware wise they
> look identical. Is there anything on the hardware side that makes the
> ACX5048 better than a QFX5100, or is it only software related?
> _______________________________________________
> juniper-nsp mailing list juniper-***@puck.nether.net
> https://puck.nether.net/mailman/listinfo/juniper-nsp
_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
https://puck.nether.net/mailman/lis

[Aaron Gould]

Not sure what Alexandre is talking about with "vlan leak"

I have bought lots of ACX5048 (about 50 of them) and use them as MPLS PE edge... For business Ethernet, cell backhaul, residential broadband, ftth, DSL, cable modem... Mpls l3vpn. Lit up some of the 40 gig interfaces lately too.... Very nice.

We have the newer ACX5448 in the lab but haven't tested with it yet. (48) 1/10 gig and (4) multi speed 10/25/40/100gig

I understand difference with ACX5048 and QFX5100 is that ACX has mpls edge stuff, and QFX has data center virtual chassis stuff. (I don't know all the details, but one is for metro-e mpls agg and the other is for DC)

Aaron
_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
https://puck.nether.net/mailman/listinfo/juniper-nsp

[Giuliano C. Medalha]

Aaron

Witch version of Junos are you running in your ACX5048 boxes ?

Is it possible to share it ?

Thanks

Giuliano

________________________________
From: juniper-nsp <juniper-nsp-***@puck.nether.net> on behalf of Aaron Gould <***@gvtc.com>
Sent: Monday, July 2, 2018 1:44:59 AM
To: Alexandre Guimaraes
Cc: Juniper List
Subject: Re: [j-nsp] QFX5100 vs ACX5048

Not sure what Alexandre is talking about with "vlan leak"

I have bought lots of ACX5048 (about 50 of them) and use them as MPLS PE edge... For business Ethernet, cell backhaul, residential broadband, ftth, DSL, cable modem... Mpls l3vpn. Lit up some of the 40 gig interfaces lately too.... Very nice.

We have the newer ACX5448 in the lab but haven't tested with it yet. (48) 1/10 gig and (4) multi speed 10/25/40/100gig

I understand difference with ACX5048 and QFX5100 is that ACX has mpls edge stuff, and QFX has data center virtual chassis stuff. (I don't know all the details, but one is for metro-e mpls agg and the other is for DC)

Aaron
_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
https://puck.nether.net/mailman/listinfo/juniper-nsp

WZTECH is registered trademark of WZTECH NETWORKS.
Copyright © 2017 WZTECH NETWORKS. All Rights Reserved.

IMPORTANTE:
As informações deste e-mail e o conteúdo dos eventuais documentos anexos são confidenciais e para conhecimento exclusivo do destinatário. Se o leitor desta mensagem não for o seu destinatário, fica desde já notificado de que não poderá divulgar, distribuir ou, sob qualquer forma, dar conhecimento a terceiros das informações e do conteúdo dos documentos anexos. Neste caso, favor comunicar imediatamente o remetente, respondendo este e-mail ou telefonando ao mesmo, e em seguida apague-o.

CONFIDENTIALITY NOTICE:
The information transmitted in this email message and any attachments are solely for the intended recipient and may contain confidential or privileged information. If you are not the intended recipient, any review, transmission, dissemination or other use of this information is prohibited. If you have received this communication in error, please notify the sender immediately and delete the material from any computer, including any copies.

WZTECH is registered trademark of WZTECH NETWORKS.
Copyright © 2017 WZTECH NETWORKS. All Rights Reserved.

IMPORTANTE:
As informações deste e-mail e o conteúdo dos eventuais documentos anexos são confidenciais e para conhecimento exclusivo do destinatário. Se o leitor desta mensagem não for o seu destinatário, fica desde já notificado de que não poderá divulgar, distribuir ou, sob qualquer forma, dar conhecimento a terceiros das informações e do conteúdo dos documentos anexos. Neste caso, favor comunicar imediatamente o remetente, respondendo este e-mail ou telefonando ao mesmo, e em seguida apague-o.

CONFIDENTIALITY NOTICE:
The information transmitted in this email message and any attachments are solely for the intended recipient and may contain confidential or privileged information. If you are not the intended recipient, any review, transmission, dissemination or other use of this information is prohibited. If you have received this communication in error, please notify the sender immediately and delete the material from any computer, including any copies.
_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
https://puck.nether.net/mailman/listinfo/juniper-nsp

[Aaron Gould]

ACX5048



15.1X54-D61.6



- Aaron



_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
https://puck.nether.net/mailman/listinfo/juniper-nsp

[Alexandre Guimaraes]

Aaron,
I am using 15.1X54-D51.7, I will try D61.6 version, Let´s see.

I moved out all Carriers NNI almost 2 years ago and never change anymore since that fatidic day, now the impact will affect only ethernet l2circuits.



--- JUNOS 15.1X54-D51.7 built 2016-08-22 17:14:59 UTC
{master:0}
> show system uptime
fpc0:
--------------------------------------------------------------------------
Current time: 2018-07-02 11:03:15 BRT
.
.
.
11:03AM up 647 days, 8:57, 1 user, load averages: 0.13, 0.10, 0.08


Att
Alexandre

De: Aaron Gould <***@gvtc.com>
Enviada em: segunda-feira, 2 de julho de 2018 10:17
Para: 'Giuliano C. Medalha' <***@wztech.com.br>; Alexandre Guimaraes <***@ascenty.com>
Cc: 'Juniper List' <juniper-***@puck.nether.net>
Assunto: RE: [j-nsp] QFX5100 vs ACX5048

ACX5048

15.1X54-D61.6

- Aaron


_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
https://puck.nether.net/mailman/listinfo/juniper-nsp

[Gustavo Santos]

Alexandre,

The 15.1X54-D67.2 Fixed this issue for Us, and when I first got this
issue, they give us a "workaround" that disabled the default arp policer
and solved the l2 circuits issues with arp packets.

On Mon, Jul 2, 2018 at 11:06 AM Alexandre Guimaraes <
***@ascenty.com> wrote:

> Aaron,
> I am using 15.1X54-D51.7, I will try D61.6 version, Let´s see.
>
> I moved out all Carriers NNI almost 2 years ago and never
> change anymore since that fatidic day, now the impact will affect only
> ethernet l2circuits.
>
>
>
> --- JUNOS 15.1X54-D51.7 built 2016-08-22 17:14:59 UTC
> {master:0}
> > show system uptime
> fpc0:
> --------------------------------------------------------------------------
> Current time: 2018-07-02 11:03:15 BRT
> .
> .
> .
> 11:03AM up 647 days, 8:57, 1 user, load averages: 0.13, 0.10, 0.08
>
>
> Att
> Alexandre
>
> De: Aaron Gould <***@gvtc.com>
> Enviada em: segunda-feira, 2 de julho de 2018 10:17
> Para: 'Giuliano C. Medalha' <***@wztech.com.br>; Alexandre Guimaraes
> <***@ascenty.com>
> Cc: 'Juniper List' <juniper-***@puck.nether.net>
> Assunto: RE: [j-nsp] QFX5100 vs ACX5048
>
> ACX5048
>
> 15.1X54-D61.6
>
> - Aaron
>
>
> _______________________________________________
> juniper-nsp mailing list juniper-***@puck.nether.net
> https://puck.nether.net/mailman/listinfo/juniper-nsp
>
_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
https://puck.nether.net/mailman/listinfo/juniper-ns

[Aaron Gould]

I think I was originally on D20, and had the arp issue… PR1271100 … https://prsearch.juniper.net/InfoCenter/index?page=prcontent <https://prsearch.juniper.net/InfoCenter/index?page=prcontent&id=PR1271100> &id=PR1271100

--- I think arp issue is seen in queue 27 (start shell…. Cprod command to see it)



I think we went to D51 to solve arp issue. I recall D51 killing my ability to do dhcp-relay. We do dhcp-relay for residential broadband… pretty sure D51 testing was great UNTIL we saw dhcp-relay broken.



We went to D61 and that’s where we sit now.



What is “vlan-leaking” ? Do you have a Juniper PR describing it please ?



I look into a preso I had and also this site… https://packetpushers.net/juniper-enterprise-serious-campus-networking/

…and I see mention of the chip for the ACX5448 possibly being Qumran-based. Not sure if that helps y’all.



I did also see on a juniper.net spec sheet that throughput for acx5048 is 1.44 tbps and acx5448 is 800 mbps. I’m wondering why is the ACX5448 less than ACX5048. I thought ACX5448 was supposed to be better, next iteration of ACX5048/96.



-Aaron





From: Gustavo Santos [mailto:***@gmail.com]
Sent: Monday, July 2, 2018 9:09 AM
To: ***@ascenty.com
Cc: Aaron Gould; Giuliano Medalha; juniper-***@puck.nether.net
Subject: Re: [j-nsp] RES: QFX5100 vs ACX5048



Alexandre,



The 15.1X54-D67.2 Fixed this issue for Us, and when I first got this issue, they give us a "workaround" that disabled the default arp policer and solved the l2 circuits issues with arp packets.



On Mon, Jul 2, 2018 at 11:06 AM Alexandre Guimaraes <***@ascenty.com> wrote:

Aaron,
I am using 15.1X54-D51.7, I will try D61.6 version, Let´s see.

I moved out all Carriers NNI almost 2 years ago and never change anymore since that fatidic day, now the impact will affect only ethernet l2circuits.



--- JUNOS 15.1X54-D51.7 built 2016-08-22 17:14:59 UTC
{master:0}
> show system uptime
fpc0:
--------------------------------------------------------------------------
Current time: 2018-07-02 11:03:15 BRT
.
.
.
11:03AM up 647 days, 8:57, 1 user, load averages: 0.13, 0.10, 0.08


Att
Alexandre

De: Aaron Gould <***@gvtc.com>
Enviada em: segunda-feira, 2 de julho de 2018 10:17
Para: 'Giuliano C. Medalha' <***@wztech.com.br>; Alexandre Guimaraes <***@ascenty.com>
Cc: 'Juniper List' <juniper-***@puck.nether.net>
Assunto: RE: [j-nsp] QFX5100 vs ACX5048

ACX5048

15.1X54-D61.6

- Aaron


_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
https://puck.nether.net/mailman/listinfo/juniper-nsp

_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
https://puck.nether.net/mailman/listinfo/j

[Luis Balbinot]

> I look into a preso I had and also this site… https://packetpushers.net/juniper-enterprise-serious-campus-networking/
>
> …and I see mention of the chip for the ACX5448 possibly being Qumran-based. Not sure if that helps y’all.

Yes, it is Qumran-based. 1M FIB, deep buffers, HQoS.

*Sounds* promising, but only if it comes out on a similar price tag as
the ACX5048. Or else there's no good reason not to use the MX204 for
PE like someone else mentioned.

There's also the ACX6030 on the horizon, running a "PTX" silicon with
a reduced feature set.

Luis
_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
https://puck.nether

[Mark Tinka]

On 2/Jul/18 01:04, Alexandre Guimaraes wrote:

>
> The new promise is the new acx5448. No vlan leaking, a good load balance(ae) algorithm, full of this.... full of that.... a lot of promise.

What chip do they have in there?

If it's still Broadcom, I'll pass...

Mark.
_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
https://puck.nether.net/mailman/listinfo/juniper-nsp

[Duane Grant]

What if it’s Jericho? Then you have something similar to the 5k ncs boxes?

Have you tested Jericho and found significant limitations?


/Duane


/Duane


> On Jul 2, 2018, at 4:35 AM, Mark Tinka <***@seacom.mu> wrote:
>
>
>
>> On 2/Jul/18 01:04, Alexandre Guimaraes wrote:
>>
>>
>> The new promise is the new acx5448. No vlan leaking, a good load balance(ae) algorithm, full of this.... full of that.... a lot of promise.
>
> What chip do they have in there?
>
> If it's still Broadcom, I'll pass...
>
> Mark.
> _______________________________________________
> juniper-nsp mailing list juniper-***@puck.nether.net
> https://puck.nether.net/mailman/listinfo/juniper-nsp
_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
http

[Mark Tinka]

On 2/Jul/18 13:55, Duane Grant wrote:

> What if it’s Jericho? Then you have something similar to the 5k ncs boxes?
>
> Have you tested Jericho and found significant limitations?

The NCS5000 failed our pre-test checklist from Cisco several months back.

I have to admit, I don't keep track of Broadcom's chips and what is
supported with each release, as I'm just too lazy, if I'm honest. Yes, I
do pay a bit of a premium for in-house chips, but the peace of mind is
worth the extra cost, over the course of the lifespan of the gear.

Which is not to say we'd never use Broadcom in the network - for
example, Layer 2 core switching or data centre aggregation is more than
fine on the Aristas we deploy. But when we need to enable IP or MPLS,
the hit & miss with Broadcom is just headache we don't need.

Mark.
_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
https://puck.nether.net/mailman/listinfo/juniper-

[Gert Doering]

This post might be inappropriate. Click to display it.

[Gustavo Santos]

I had some issues with ACX5048 , the most noticable was arp packets from
pure L2 vlans and VPLS punting to CPU and flooding the default arp policer
police..
JTAC was able to reproduce the issue and gave us an option to disable
default arp policer until they release a service release to fix this issue
that was solved indeed.

Another problem was upgrading to the lastest Junos JTAC recommended that
made the ACX5048 unusable... ( Junos was unable to find the physical
ports..) We had to downgrade to get it back working again..

On Sun, Jul 1, 2018 at 8:05 PM Alexandre Guimaraes <
***@ascenty.com> wrote:

> Better in terms of concept. In term of usage, i still investing in qfx5100
>
> Acx5058 Suppose to be a promise of a new future, unfortunately, with all
> problematic of the qfx5100 hardware, the acx5048 leak vlan till the last
> breath of cpu.... after that, all deamons and services going down.... up
> and down, up and down.
>
> I never more brought one peace ACX5048 after jtac didnt responds why and
> solution for the leaking...( I have only two acx5048 and hundreds on
> QFX...).
>
> The new promise is the new acx5448. No vlan leaking, a good load
> balance(ae) algorithm, full of this.... full of that.... a lot of promise.
>
> Let’s see...
>
> att
> Alexandre
>
> Em 1 de jul de 2018, à(s) 19:31, Colton Conor <***@gmail.com>
> escreveu:
>
> > What is the main difference between these two boxes? Hardware wise they
> > look identical. Is there anything on the hardware side that makes the
> > ACX5048 better than a QFX5100, or is it only software related?
> > _______________________________________________
> > juniper-nsp mailing list juniper-***@puck.nether.net
> > https://puck.nether.net/mailman/listinfo/juniper-nsp
> _______________________________________________
> juniper-nsp mailing list juniper-***@puck.nether.net
> https://puck.nether.net/mailman/listinfo/juniper-nsp
>
_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
https://puck.nethe

[Mark Tinka]

On 2/Jul/18 15:11, Gustavo Santos wrote:

> I had some issues with ACX5048 , the most noticable was arp packets from
> pure L2 vlans and VPLS punting to CPU and flooding the default arp policer
> police..
> JTAC was able to reproduce the issue and gave us an option to disable
> default arp policer until they release a service release to fix this issue
> that was solved indeed.
>
> Another problem was upgrading to the lastest Junos JTAC recommended that
> made the ACX5048 unusable... ( Junos was unable to find the physical
> ports..) We had to downgrade to get it back working again..

Since 2008, Juniper have cocked up every opportunity to release a solid
Metro-E box that can compete with the Cisco and Brocade options.

As big as it can be seen to be, the MX204 is probably Juniper's first
real attempt at this.

Mark.
_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
https://puck.nether.net/mailman/listinfo/juniper-nsp

[Alexandre Guimaraes]

Same over here.... vlan leaking

De: Gustavo Santos <***@gmail.com>
Enviada em: segunda-feira, 2 de julho de 2018 10:11
Para: Alexandre Guimaraes <***@ascenty.com>
Cc: Colton Conor <***@gmail.com>; juniper-***@puck.nether.net
Assunto: Re: [j-nsp] QFX5100 vs ACX5048

I had some issues with ACX5048 , the most noticable was arp packets from pure L2 vlans and VPLS punting to CPU and flooding the default arp policer police..
JTAC was able to reproduce the issue and gave us an option to disable default arp policer until they release a service release to fix this issue that was solved indeed.

Another problem was upgrading to the lastest Junos JTAC recommended that made the ACX5048 unusable... ( Junos was unable to find the physical ports..) We had to downgrade to get it back working again..

On Sun, Jul 1, 2018 at 8:05 PM Alexandre Guimaraes <***@ascenty.com<mailto:***@ascenty.com>> wrote:
Better in terms of concept. In term of usage, i still investing in qfx5100

Acx5058 Suppose to be a promise of a new future, unfortunately, with all problematic of the qfx5100 hardware, the acx5048 leak vlan till the last breath of cpu.... after that, all deamons and services going down.... up and down, up and down.

I never more brought one peace ACX5048 after jtac didnt responds why and solution for the leaking...( I have only two acx5048 and hundreds on QFX...).

The new promise is the new acx5448. No vlan leaking, a good load balance(ae) algorithm, full of this.... full of that.... a lot of promise.

Let’s see...

att
Alexandre

Em 1 de jul de 2018, à(s) 19:31, Colton Conor <***@gmail.com<mailto:***@gmail.com>> escreveu:

> What is the main difference between these two boxes? Hardware wise they
> look identical. Is there anything on the hardware side that makes the
> ACX5048 better than a QFX5100, or is it only software related?
> _______________________________________________
> juniper-nsp mailing list juniper-***@puck.nether.net<mailto:juniper-***@puck.nether.net>
> https://puck.nether.net/mailman/listinfo/juniper-nsp
_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net<mailto:juniper-***@puck.nether.net>
https://puck.nether.net/mailman/listinfo/juniper-nsp
_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
https://puck.nether.net/mailman/listinfo/juniper-nsp

[Colton Conor]

Gustavo,

We you say " Another problem was upgrading to the lastest Junos JTAC
recommended that made the ACX5048 unusable... ( Junos was unable to find
the physical ports..) We had to downgrade to get it back working again.."
what version was this as JTAC recently changed their recommended version?
It seem everyone on this thread is talking about software train 15.1X54.

However, the current JTAC recommended version is Junos 17.3R2 as of 14 May
2018. Why is everyone running 15.1X54 code?

Has anyone upgraded to Junos 17.3R2 on the ACX's? No matter the model, all
ACX's current recommended is now Junos 17.3R2


On Mon, Jul 2, 2018 at 8:11 AM, Gustavo Santos <***@gmail.com> wrote:

> I had some issues with ACX5048 , the most noticable was arp packets from
> pure L2 vlans and VPLS punting to CPU and flooding the default arp policer
> police..
> JTAC was able to reproduce the issue and gave us an option to disable
> default arp policer until they release a service release to fix this issue
> that was solved indeed.
>
> Another problem was upgrading to the lastest Junos JTAC recommended that
> made the ACX5048 unusable... ( Junos was unable to find the physical
> ports..) We had to downgrade to get it back working again..
>
> On Sun, Jul 1, 2018 at 8:05 PM Alexandre Guimaraes <
> ***@ascenty.com> wrote:
>
>> Better in terms of concept. In term of usage, i still investing in qfx5100
>>
>> Acx5058 Suppose to be a promise of a new future, unfortunately, with all
>> problematic of the qfx5100 hardware, the acx5048 leak vlan till the last
>> breath of cpu.... after that, all deamons and services going down.... up
>> and down, up and down.
>>
>> I never more brought one peace ACX5048 after jtac didnt responds why and
>> solution for the leaking...( I have only two acx5048 and hundreds on
>> QFX...).
>>
>> The new promise is the new acx5448. No vlan leaking, a good load
>> balance(ae) algorithm, full of this.... full of that.... a lot of promise.
>>
>> Let’s see...
>>
>> att
>> Alexandre
>>
>> Em 1 de jul de 2018, à(s) 19:31, Colton Conor <***@gmail.com>
>> escreveu:
>>
>> > What is the main difference between these two boxes? Hardware wise they
>> > look identical. Is there anything on the hardware side that makes the
>> > ACX5048 better than a QFX5100, or is it only software related?
>> > _______________________________________________
>> > juniper-nsp mailing list juniper-***@puck.nether.net
>> > https://puck.nether.net/mailman/listinfo/juniper-nsp
>> _______________________________________________
>> juniper-nsp mailing list juniper-***@puck.nether.net
>> https://puck.nether.net/mailman/listinfo/juniper-nsp
>>
>
_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net

[Aaron Gould]

I don't upgrade software unless I have a good reason to...especially if my network is fairly stable and I know what I have.

Aaron

> On Jul 4, 2018, at 9:31 AM, Colton Conor <***@gmail.com> wrote:
>
> Gustavo,
>
> We you say " Another problem was upgrading to the lastest Junos JTAC recommended that made the ACX5048 unusable... ( Junos was unable to find the physical ports..) We had to downgrade to get it back working again.." what version was this as JTAC recently changed their recommended version? It seem everyone on this thread is talking about software train 15.1X54.
>
> However, the current JTAC recommended version is Junos 17.3R2 as of 14 May 2018. Why is everyone running 15.1X54 code?
>
> Has anyone upgraded to Junos 17.3R2 on the ACX's? No matter the model, all ACX's current recommended is now Junos 17.3R2
>
>
>> On Mon, Jul 2, 2018 at 8:11 AM, Gustavo Santos <***@gmail.com> wrote:
>> I had some issues with ACX5048 , the most noticable was arp packets from pure L2 vlans and VPLS punting to CPU and flooding the default arp policer police..
>> JTAC was able to reproduce the issue and gave us an option to disable default arp policer until they release a service release to fix this issue that was solved indeed.
>>
>> Another problem was upgrading to the lastest Junos JTAC recommended that made the ACX5048 unusable... ( Junos was unable to find the physical ports..) We had to downgrade to get it back working again..
>>
>>> On Sun, Jul 1, 2018 at 8:05 PM Alexandre Guimaraes <***@ascenty.com> wrote:
>>> Better in terms of concept. In term of usage, i still investing in qfx5100
>>>
>>> Acx5058 Suppose to be a promise of a new future, unfortunately, with all problematic of the qfx5100 hardware, the acx5048 leak vlan till the last breath of cpu.... after that, all deamons and services going down.... up and down, up and down.
>>>
>>> I never more brought one peace ACX5048 after jtac didnt responds why and solution for the leaking...( I have only two acx5048 and hundreds on QFX...).
>>>
>>> The new promise is the new acx5448. No vlan leaking, a good load balance(ae) algorithm, full of this.... full of that.... a lot of promise.
>>>
>>> Let’s see...
>>>
>>> att
>>> Alexandre
>>>
>>> Em 1 de jul de 2018, à(s) 19:31, Colton Conor <***@gmail.com> escreveu:
>>>
>>> > What is the main difference between these two boxes? Hardware wise they
>>> > look identical. Is there anything on the hardware side that makes the
>>> > ACX5048 better than a QFX5100, or is it only software related?
>>> > _______________________________________________
>>> > juniper-nsp mailing list juniper-***@puck.nether.net
>>> > https://puck.nether.net/mailman/listinfo/juniper-nsp
>>> _______________________________________________
>>> juniper-nsp mailing list juniper-***@puck.nether.net
>>> https://puck.nether.net/mailman/listinfo/juniper-nsp
>
_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.

[Colton Conor]

Nick,

Did you find the PR for this memory leak?

On Wed, Jul 4, 2018 at 11:02 AM, Nick Ryce <***@commsworld.com> wrote:

> If you use BFD, do not upgrade to 17.3R2 as there is a memory leak. Will
> find the PR.
>
> N
>
> On 04/07/2018, 15:31, "juniper-nsp on behalf of Colton Conor" <
> juniper-nsp-***@puck.nether.net on behalf of ***@gmail.com>
> wrote:
>
> Gustavo,
>
> We you say " Another problem was upgrading to the lastest Junos JTAC
> recommended that made the ACX5048 unusable... ( Junos was unable to
> find
> the physical ports..) We had to downgrade to get it back working
> again.."
> what version was this as JTAC recently changed their recommended
> version?
> It seem everyone on this thread is talking about software train
> 15.1X54.
>
> However, the current JTAC recommended version is Junos 17.3R2 as of 14
> May
> 2018. Why is everyone running 15.1X54 code?
>
> Has anyone upgraded to Junos 17.3R2 on the ACX's? No matter the
> model, all
> ACX's current recommended is now Junos 17.3R2
>
>
> On Mon, Jul 2, 2018 at 8:11 AM, Gustavo Santos <***@gmail.com>
> wrote:
>
> > I had some issues with ACX5048 , the most noticable was arp packets
> from
> > pure L2 vlans and VPLS punting to CPU and flooding the default arp
> policer
> > police..
> > JTAC was able to reproduce the issue and gave us an option to disable
> > default arp policer until they release a service release to fix this
> issue
> > that was solved indeed.
> >
> > Another problem was upgrading to the lastest Junos JTAC recommended
> that
> > made the ACX5048 unusable... ( Junos was unable to find the physical
> > ports..) We had to downgrade to get it back working again..
> >
> > On Sun, Jul 1, 2018 at 8:05 PM Alexandre Guimaraes <
> > ***@ascenty.com> wrote:
> >
> >> Better in terms of concept. In term of usage, i still investing in
> qfx5100
> >>
> >> Acx5058 Suppose to be a promise of a new future, unfortunately,
> with all
> >> problematic of the qfx5100 hardware, the acx5048 leak vlan till the
> last
> >> breath of cpu.... after that, all deamons and services going
> down.... up
> >> and down, up and down.
> >>
> >> I never more brought one peace ACX5048 after jtac didnt responds
> why and
> >> solution for the leaking...( I have only two acx5048 and hundreds on
> >> QFX...).
> >>
> >> The new promise is the new acx5448. No vlan leaking, a good load
> >> balance(ae) algorithm, full of this.... full of that.... a lot of
> promise.
> >>
> >> Let’s see...
> >>
> >> att
> >> Alexandre
> >>
> >> Em 1 de jul de 2018, à(s) 19:31, Colton Conor <
> ***@gmail.com>
> >> escreveu:
> >>
> >> > What is the main difference between these two boxes? Hardware
> wise they
> >> > look identical. Is there anything on the hardware side that makes
> the
> >> > ACX5048 better than a QFX5100, or is it only software related?
> >> > _______________________________________________
> >> > juniper-nsp mailing list juniper-***@puck.nether.net
> >> > https://puck.nether.net/mailman/listinfo/juniper-nsp
> >> _______________________________________________
> >> juniper-nsp mailing list juniper-***@puck.nether.net
> >> https://puck.nether.net/mailman/listinfo/juniper-nsp
> >>
> >
> _______________________________________________
> juniper-nsp mailing list juniper-***@puck.nether.net
> https://puck.nether.net/mailman/listinfo/juniper-nsp
>
>
>
_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
https://puck.n

[Colton Conor]

Nick,

Why does it say
*Resolved In* 16.2R1-S7 16.2R3 17.1R2 17.1R3 17.2R2 *17.3R1*
17.3R1 came before the now current JTAC recommend *17.3R2 right? *

On Wed, Jul 11, 2018 at 6:39 AM, Nick Ryce <***@commsworld.com> wrote:

> Sorry I thought I had.
>
>
>
> We hit this https://prsearch.juniper.net/InfoCenter/index?page=
> prcontent&id=PR1280492
>
>
>
> This is being resolved in 17.3R3 which is due for release in a couple of
> weeks.
>
>
>
> N
>
>
>
> *From: *Colton Conor <***@gmail.com>
> *Date: *Wednesday, 11 July 2018 at 12:30
> *To: *Nick Ryce <***@commsworld.com>
> *Cc: *Gustavo Santos <***@gmail.com>, Aaron <***@gvtc.com>,
> Juniper List <juniper-***@puck.nether.net>
> *Subject: *Re: [j-nsp] QFX5100 vs ACX5048
>
>
>
> Nick,
>
>
>
> Did you find the PR for this memory leak?
>
>
>
> On Wed, Jul 4, 2018 at 11:02 AM, Nick Ryce <***@commsworld.com>
> wrote:
>
> If you use BFD, do not upgrade to 17.3R2 as there is a memory leak. Will
> find the PR.
>
> N
>
>
> On 04/07/2018, 15:31, "juniper-nsp on behalf of Colton Conor" <
> juniper-nsp-***@puck.nether.net on behalf of ***@gmail.com>
> wrote:
>
> Gustavo,
>
> We you say " Another problem was upgrading to the lastest Junos JTAC
> recommended that made the ACX5048 unusable... ( Junos was unable to
> find
> the physical ports..) We had to downgrade to get it back working
> again.."
> what version was this as JTAC recently changed their recommended
> version?
> It seem everyone on this thread is talking about software train
> 15.1X54.
>
> However, the current JTAC recommended version is Junos 17.3R2 as of 14
> May
> 2018. Why is everyone running 15.1X54 code?
>
> Has anyone upgraded to Junos 17.3R2 on the ACX's? No matter the
> model, all
> ACX's current recommended is now Junos 17.3R2
>
>
> On Mon, Jul 2, 2018 at 8:11 AM, Gustavo Santos <***@gmail.com>
> wrote:
>
> > I had some issues with ACX5048 , the most noticable was arp packets
> from
> > pure L2 vlans and VPLS punting to CPU and flooding the default arp
> policer
> > police..
> > JTAC was able to reproduce the issue and gave us an option to disable
> > default arp policer until they release a service release to fix this
> issue
> > that was solved indeed.
> >
> > Another problem was upgrading to the lastest Junos JTAC recommended
> that
> > made the ACX5048 unusable... ( Junos was unable to find the physical
> > ports..) We had to downgrade to get it back working again..
> >
> > On Sun, Jul 1, 2018 at 8:05 PM Alexandre Guimaraes <
> > ***@ascenty.com> wrote:
> >
> >> Better in terms of concept. In term of usage, i still investing in
> qfx5100
> >>
> >> Acx5058 Suppose to be a promise of a new future, unfortunately,
> with all
> >> problematic of the qfx5100 hardware, the acx5048 leak vlan till the
> last
> >> breath of cpu.... after that, all deamons and services going
> down.... up
> >> and down, up and down.
> >>
> >> I never more brought one peace ACX5048 after jtac didnt responds
> why and
> >> solution for the leaking...( I have only two acx5048 and hundreds on
> >> QFX...).
> >>
> >> The new promise is the new acx5448. No vlan leaking, a good load
> >> balance(ae) algorithm, full of this.... full of that.... a lot of
> promise.
> >>
> >> Let’s see...
> >>
> >> att
> >> Alexandre
> >>
> >> Em 1 de jul de 2018, à(s) 19:31, Colton Conor <
> ***@gmail.com>
> >> escreveu:
> >>
> >> > What is the main difference between these two boxes? Hardware
> wise they
> >> > look identical. Is there anything on the hardware side that makes
> the
> >> > ACX5048 better than a QFX5100, or is it only software related?
> >> > _______________________________________________
> >> > juniper-nsp mailing list juniper-***@puck.nether.net
> >> > https://puck.nether.net/mailman/listinfo/juniper-nsp
> >> _______________________________________________
> >> juniper-nsp mailing list juniper-***@puck.nether.net
> >> https://puck.nether.net/mailman/listinfo/juniper-nsp
> >>
> >
> _______________________________________________
> juniper-nsp mailing list juniper-***@puck.nether.net
> https://puck.nether.net/mailman/listinfo/juniper-nsp
>
>
>
_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
https://puck.nether.net/mailman/

[Jackson, William]

> ***@gmail.com>
> wrote:
>
> Gustavo,
>
> We you say " Another problem was upgrading to the lastest Junos JTAC
> recommended that made the ACX5048 unusable... ( Junos was unable
> to find
> the physical ports..) We had to downgrade to get it back working
> again.."

How the hell does that type of problem even get past QA?

I had the same issue with a qfx5100, upgraded to 17.x version and when booted didn’t detect the ports. After a 30 minute wait we downgraded.

Insane.
_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
ht

[Aaron Gould]

I was wondering the same thing in 15.1X54-D51.7 with DHCP-relay on IRB's not working inside my L3VPN , I mean I was wondering how did that pass internally testing before it was released. I'm asking that not knowing anything about how Juniper tests their boxes and code revs... But simply thinking that if I can find a problem, surely they could have and just wondering, like you are, why they didn't

Aaron

On Jul 11, 2018, at 7:28 AM, Jackson, William <***@gibtele.com> wrote:

>> ***@gmail.com>
>> wrote:
>>
>> Gustavo,
>>
>> We you say " Another problem was upgrading to the lastest Junos JTAC
>> recommended that made the ACX5048 unusable... ( Junos was unable
>> to find
>> the physical ports..) We had to downgrade to get it back working
>> again.."
>
> How the hell does that type of problem even get past QA?
>
> I had the same issue with a qfx5100, upgraded to 17.x version and when booted didn’t detect the ports. After a 30 minute wait we downgraded.
>
> Insane.
> _______________________________________________
> juniper-nsp mailing list juniper-***@puck.nether.net
> https://puck.nether.net/mailman/listinfo/juniper-nsp

_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
https://puck.nether.net/mail

[a***@netconsultings.com]

> Of Jackson, William
> Sent: Wednesday, July 11, 2018 1:28 PM
> To: 'Colton Conor'; Nick Ryce
> Cc: Juniper List
> Subject: Re: [j-nsp] QFX5100 vs ACX5048
>
> > ***@gmail.com>
> > wrote:
> >
> > Gustavo,
> >
> > We you say " Another problem was upgrading to the lastest Junos JTAC
> > recommended that made the ACX5048 unusable... ( Junos was unable
> > to find
> > the physical ports..) We had to downgrade to get it back working
> > again.."
>
> How the hell does that type of problem even get past QA?
>
> I had the same issue with a qfx5100, upgraded to 17.x version and when
> booted didn’t detect the ports. After a 30 minute wait we downgraded.
>
We had the same exact fun with the good old ME3600, makes you wonder whether vendors have time to do any testing when in pressure to release a new code.

adam

netconsultings.com
::carrier-class solutions for the telecommunications industry::


_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
https://puck.nether.net/mailman/listinfo/junipe

[Aaron Gould]

Right.

Growing pains I guess. Is anything every bug free? I think the more features you put into something the more possibility there is for bugs.

Actually I feel this is why when I get comfortable with a version of code and the platform I tend to camp out there for quite some time and have to be really convinced to load any new software on it.

I remember hitting a IPv4 forwarding bug in IOS XR 4.something on my ASR9k around 10:00 a.m. Some years as back.... Dang I actually remember the time of day... Traumatized I guess , LOL , I remember my boss and everyone standing over me in my office while we worked frantically with the XR TAC about loading a fix

Aaron

On Jul 11, 2018, at 8:14 AM, <***@netconsultings.com> <***@netconsultings.com> wrote:

>> Of Jackson, William
>> Sent: Wednesday, July 11, 2018 1:28 PM
>> To: 'Colton Conor'; Nick Ryce
>> Cc: Juniper List
>> Subject: Re: [j-nsp] QFX5100 vs ACX5048
>>
>>> ***@gmail.com>
>>> wrote:
>>>
>>> Gustavo,
>>>
>>> We you say " Another problem was upgrading to the lastest Junos JTAC
>>> recommended that made the ACX5048 unusable... ( Junos was unable
>>> to find
>>> the physical ports..) We had to downgrade to get it back working
>>> again.."
>>
>> How the hell does that type of problem even get past QA?
>>
>> I had the same issue with a qfx5100, upgraded to 17.x version and when
>> booted didn’t detect the ports. After a 30 minute wait we downgraded.
> We had the same exact fun with the good old ME3600, makes you wonder whether vendors have time to do any testing when in pressure to release a new code.
>
> adam
>
> netconsultings.com
> ::carrier-class solutions for the telecommunications industry::
>
>
> _______________________________________________
> juniper-nsp mailing list juniper-***@puck.nether.net
> https://puck.nether.net/mailman/listinfo/juniper-nsp

_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net

[Gustavo Santos]

Yes... Exactly.

I upgraded to 17.3R2 and was a total mess. And had to downgrade to 15.1X
train.


On Wed, Jul 4, 2018 at 11:31 AM Colton Conor <***@gmail.com> wrote:

> Gustavo,
>
> We you say " Another problem was upgrading to the lastest Junos JTAC
> recommended that made the ACX5048 unusable... ( Junos was unable to find
> the physical ports..) We had to downgrade to get it back working again.."
> what version was this as JTAC recently changed their recommended version?
> It seem everyone on this thread is talking about software train 15.1X54.
>
> However, the current JTAC recommended version is Junos 17.3R2 as of 14 May
> 2018. Why is everyone running 15.1X54 code?
>
> Has anyone upgraded to Junos 17.3R2 on the ACX's? No matter the model,
> all ACX's current recommended is now Junos 17.3R2
>
>
> On Mon, Jul 2, 2018 at 8:11 AM, Gustavo Santos <***@gmail.com>
> wrote:
>
>> I had some issues with ACX5048 , the most noticable was arp packets from
>> pure L2 vlans and VPLS punting to CPU and flooding the default arp policer
>> police..
>> JTAC was able to reproduce the issue and gave us an option to disable
>> default arp policer until they release a service release to fix this issue
>> that was solved indeed.
>>
>> Another problem was upgrading to the lastest Junos JTAC recommended that
>> made the ACX5048 unusable... ( Junos was unable to find the physical
>> ports..) We had to downgrade to get it back working again..
>>
>> On Sun, Jul 1, 2018 at 8:05 PM Alexandre Guimaraes <
>> ***@ascenty.com> wrote:
>>
>>> Better in terms of concept. In term of usage, i still investing in
>>> qfx5100
>>>
>>> Acx5058 Suppose to be a promise of a new future, unfortunately, with all
>>> problematic of the qfx5100 hardware, the acx5048 leak vlan till the last
>>> breath of cpu.... after that, all deamons and services going down.... up
>>> and down, up and down.
>>>
>>> I never more brought one peace ACX5048 after jtac didnt responds why and
>>> solution for the leaking...( I have only two acx5048 and hundreds on
>>> QFX...).
>>>
>>> The new promise is the new acx5448. No vlan leaking, a good load
>>> balance(ae) algorithm, full of this.... full of that.... a lot of promise.
>>>
>>> Let’s see...
>>>
>>> att
>>> Alexandre
>>>
>>> Em 1 de jul de 2018, à(s) 19:31, Colton Conor <***@gmail.com>
>>> escreveu:
>>>
>>> > What is the main difference between these two boxes? Hardware wise they
>>> > look identical. Is there anything on the hardware side that makes the
>>> > ACX5048 better than a QFX5100, or is it only software related?
>>> > _______________________________________________
>>> > juniper-nsp mailing list juniper-***@puck.nether.net
>>> > https://puck.nether.net/mailman/listinfo/juniper-nsp
>>> _______________________________________________
>>> juniper-nsp mailing list juniper-***@puck.nether.net
>>> https://puck.nether.net/mailman/listinfo/juniper-nsp
>>>
>>
>
_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
https://puck.nether.net/m

[Colton Conor]

Strange we had no issues upgrading from 17.3R2 from 15.1X.

On Tue, Jul 17, 2018 at 2:27 PM, Gustavo Santos <***@gmail.com>
wrote:

> Yes... Exactly.
>
> I upgraded to 17.3R2 and was a total mess. And had to downgrade to 15.1X
> train.
>
>
> On Wed, Jul 4, 2018 at 11:31 AM Colton Conor <***@gmail.com>
> wrote:
>
>> Gustavo,
>>
>> We you say " Another problem was upgrading to the lastest Junos JTAC
>> recommended that made the ACX5048 unusable... ( Junos was unable to find
>> the physical ports..) We had to downgrade to get it back working again.."
>> what version was this as JTAC recently changed their recommended version?
>> It seem everyone on this thread is talking about software train 15.1X54.
>>
>> However, the current JTAC recommended version is Junos 17.3R2 as of 14
>> May 2018. Why is everyone running 15.1X54 code?
>>
>> Has anyone upgraded to Junos 17.3R2 on the ACX's? No matter the model,
>> all ACX's current recommended is now Junos 17.3R2
>>
>>
>> On Mon, Jul 2, 2018 at 8:11 AM, Gustavo Santos <***@gmail.com>
>> wrote:
>>
>>> I had some issues with ACX5048 , the most noticable was arp packets from
>>> pure L2 vlans and VPLS punting to CPU and flooding the default arp policer
>>> police..
>>> JTAC was able to reproduce the issue and gave us an option to disable
>>> default arp policer until they release a service release to fix this issue
>>> that was solved indeed.
>>>
>>> Another problem was upgrading to the lastest Junos JTAC recommended that
>>> made the ACX5048 unusable... ( Junos was unable to find the physical
>>> ports..) We had to downgrade to get it back working again..
>>>
>>> On Sun, Jul 1, 2018 at 8:05 PM Alexandre Guimaraes <
>>> ***@ascenty.com> wrote:
>>>
>>>> Better in terms of concept. In term of usage, i still investing in
>>>> qfx5100
>>>>
>>>> Acx5058 Suppose to be a promise of a new future, unfortunately, with
>>>> all problematic of the qfx5100 hardware, the acx5048 leak vlan till the
>>>> last breath of cpu.... after that, all deamons and services going down....
>>>> up and down, up and down.
>>>>
>>>> I never more brought one peace ACX5048 after jtac didnt responds why
>>>> and solution for the leaking...( I have only two acx5048 and hundreds on
>>>> QFX...).
>>>>
>>>> The new promise is the new acx5448. No vlan leaking, a good load
>>>> balance(ae) algorithm, full of this.... full of that.... a lot of promise.
>>>>
>>>> Let’s see...
>>>>
>>>> att
>>>> Alexandre
>>>>
>>>> Em 1 de jul de 2018, à(s) 19:31, Colton Conor <***@gmail.com>
>>>> escreveu:
>>>>
>>>> > What is the main difference between these two boxes? Hardware wise
>>>> they
>>>> > look identical. Is there anything on the hardware side that makes the
>>>> > ACX5048 better than a QFX5100, or is it only software related?
>>>> > _______________________________________________
>>>> > juniper-nsp mailing list juniper-***@puck.nether.net
>>>> > https://puck.nether.net/mailman/listinfo/juniper-nsp
>>>> _______________________________________________
>>>> juniper-nsp mailing list juniper-***@puck.nether.net
>>>> https://puck.nether.net/mailman/listinfo/juniper-nsp
>>>>
>>>
>>
_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
https://puck.nether.net/ma

[Alexandre Snarskii]

On Sun, Jul 01, 2018 at 05:30:09PM -0500, Colton Conor wrote:
> What is the main difference between these two boxes? Hardware wise they
> look identical. Is there anything on the hardware side that makes the
> ACX5048 better than a QFX5100, or is it only software related?

ACX5048 uses exactly the same hardware as QFX5100, all the differences
are in the software.

_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
https://puck.nether.net/mailman/listinfo/juniper-nsp

[Aaron Gould]

Check it out... during "request system software add...." junos upgrade on
ACX5048, prior to the required reboot to make new software active, it's
interesting how there is mention of QFX :)

I think this is when I was going from D51 to D61...

...
QFX Installation Software [15.1X54-D61.6]
...


{master:0}
***@eng-lab-5048-3> show version
fpc0:
--------------------------------------------------------------------------
Hostname: eng-lab-5048-3
Model: acx5048
Junos: 15.1X54-D51.7
JUNOS Base OS boot [15.1X54-D51.7]
JUNOS Online Documentation [15.1X54-D51.7]
JUNOS Crypto Software Suite [15.1X54-D51.7]
JUNOS Base OS Software Suite [15.1X54-D51.7]
JUNOS Packet Forwarding Engine Support (acx5k) [15.1X54-D51.7]
JUNOS Kernel Software Suite [15.1X54-D51.7]
JUNOS Enterprise Software Suite [15.1X54-D51.7]
JUNOS Routing Software Suite [15.1X54-D51.7]
JUNOS py-base-i386 [15.1X54-D51.7]
QFX Installation Software [15.1X54-D61.6]
JUNOS Host Software [15.1X54-D51.7]

{master:0}


-Aaron


_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
https://puck.nether.net/mailman/listinfo/juniper-nsp

[Mark Tinka]

On 2/Jul/18 15:22, Aaron Gould wrote:
> Check it out... during "request system software add...." junos upgrade on
> ACX5048, prior to the required reboot to make new software active, it's
> interesting how there is mention of QFX :)

It's the same box, as others have stated - just different colors and
code :-).

Mark.
_______________________________________________
juniper-nsp mailing list juniper-***@puck.nether.net
https://puck.nether.net/mailman/listinfo/juniper-nsp