Switching

Hi All,

Does anyone have a working example of SRX345 using 1000Base-LX optics to Cisco 9200L switch with 1000Base-LX ? 

We have an issue with our SRX345 cluster where we had 2x 1000Base-LX interfaces towards a customer's 2x Cisco 3750 stack with Cisco 1000Base-LX optics. 

This configuration was working absolutely fine, until the customer replaced their switches with 2x Cisco C9200L-24T-4X running 17.09.04a  CAT9K_LITE_IOSXE

After reconnecting the optics, the customers Cisco SFP interfaces showed the link as UP on their side, however the SRX continues to see the interfaces as DOWN. 

Confirm we receive optical light power from the Cisco SFP, and if customer shuts down their SFP we see the optical light power drop off. Similarly if we shut our SFP, the customers SFP interface goes down, so it's not a patching issue, and keep in mind this exact link and patching was working fine, it was only an issue after the customer Cisco switches were replaced with newer model. 

Have been working through with JTAC and the suggestion from them was to ensure the customers Cisco interfaces have "speed nonegotiate" and "switchport nonegotiate" set which they do. We have similarly tried setting no-auto-negotiation on our SRX interfaces, and our side still refuses to come up, while the Cisco side is happily showing as UP and that they are sending bytes to us but not receiving any bytes back. 

I'm at a complete loss on this as to why we can't interface directly - I've tried different JunOS releases on the SRX with no difference. At the moment I have the main links back in production by sitting 2x EX2300-C-12P's in between each fibre link to act as an expensive managed SFP-to-SFP media converter, but it would be far more preferable to have the direct links working again so that the SRX345 cluster will failover between interfaces properly, as currently if the active interface were to fail between the Cisco and 2300-C the SRX on our side won't see that due to the 2300-C <-> SRX interface still being online. 

I currently have a testlab SRX345 using the Genuine Juniper SFP-1G-LX-C Part Number: 740-011614 optic towards the customers Cisco on a spare SFP port using a Genuine Cisco GLC-LH-SM 1000Base-LX optic. 

Does anyone know if there's any technical difference with the Device Specific Juniper optic "EX-SFP-1GE-LX Part Number: 740-011614 Small Form Factor Pluggable 1000Base-LX Gigabit Ethernet Optics" compared to the Common optic SFP-1G-LX-C Part Number: 740-011614 ?

Or whether there is some other optic that would be compatible, or could it genuinely be an issue with the Customer side optics in their switch as they are still using the original Cisco GLC-LH-SM modules which are quite old now.

Thanks

In these link speed configurations with Cisco in addition to setting no negotiate, we also had to explicitly configure speed 1g and duplex full for the link to come up.

------------------------------
Steve Puluka BSEET - Juniper Ambassador
IP Architect - DQE Communications Pittsburgh, PA (Metro Ethernet & ISP - Retired)
http://puluka.com/home
------------------------------

Original Message:
Sent: 08-14-2024 21:56
From: ChrisLee
Subject: SRX345 1000Base-LX Optic compatibility to Cisco 9200L

Hi All,

Does anyone have a working example of SRX345 using 1000Base-LX optics to Cisco 9200L switch with 1000Base-LX ? 

We have an issue with our SRX345 cluster where we had 2x 1000Base-LX interfaces towards a customer's 2x Cisco 3750 stack with Cisco 1000Base-LX optics. 

This configuration was working absolutely fine, until the customer replaced their switches with 2x Cisco C9200L-24T-4X running 17.09.04a  CAT9K_LITE_IOSXE

After reconnecting the optics, the customers Cisco SFP interfaces showed the link as UP on their side, however the SRX continues to see the interfaces as DOWN. 

Confirm we receive optical light power from the Cisco SFP, and if customer shuts down their SFP we see the optical light power drop off. Similarly if we shut our SFP, the customers SFP interface goes down, so it's not a patching issue, and keep in mind this exact link and patching was working fine, it was only an issue after the customer Cisco switches were replaced with newer model. 

Have been working through with JTAC and the suggestion from them was to ensure the customers Cisco interfaces have "speed nonegotiate" and "switchport nonegotiate" set which they do. We have similarly tried setting no-auto-negotiation on our SRX interfaces, and our side still refuses to come up, while the Cisco side is happily showing as UP and that they are sending bytes to us but not receiving any bytes back. 

I'm at a complete loss on this as to why we can't interface directly - I've tried different JunOS releases on the SRX with no difference. At the moment I have the main links back in production by sitting 2x EX2300-C-12P's in between each fibre link to act as an expensive managed SFP-to-SFP media converter, but it would be far more preferable to have the direct links working again so that the SRX345 cluster will failover between interfaces properly, as currently if the active interface were to fail between the Cisco and 2300-C the SRX on our side won't see that due to the 2300-C <-> SRX interface still being online. 

I currently have a testlab SRX345 using the Genuine Juniper SFP-1G-LX-C Part Number: 740-011614 optic towards the customers Cisco on a spare SFP port using a Genuine Cisco GLC-LH-SM 1000Base-LX optic. 

Does anyone know if there's any technical difference with the Device Specific Juniper optic "EX-SFP-1GE-LX Part Number: 740-011614 Small Form Factor Pluggable 1000Base-LX Gigabit Ethernet Optics" compared to the Common optic SFP-1G-LX-C Part Number: 740-011614 ?

Or whether there is some other optic that would be compatible, or could it genuinely be an issue with the Customer side optics in their switch as they are still using the original Cisco GLC-LH-SM modules which are quite old now.

Thanks