Yes, the customer used /24 subnets when he set up his network and we will fix it. Unfortunately we were not present at the initial setup and we had to deal with much bigger problems before the IP configuration.
First of all I tested TCP port 179 and it works fine. No ff is applied on lo0.0. The MTU values are mutually left at default values and the two devices are directly connected to each other. There is no reason to use Multihop, we are directly neighboring over physical interfaces. The devices are MX204 and their version is 22.4R3 (I upgraded to this version, it was version 18.2 before and I did not skip 4 versions directly while doing this, I upgraded sequentially and there was no problem). This version was also the version recommended by Juniper.
They do not communicate over any tunnel, after the BGP neighborhood is established, the GRE tunnel will pass over it for SIP traffic, but there is no such situation now.
This is what surprised me. There is no mention of any BGP bug in the version documentation, the configuration is too simple and plain to make mistakes, but I get Cease error.
Thank you very much for your help, your effort is really valuable. As you said, I cannot share more. I think I have no choice but to open the JTAC case, maybe there is a problem with one of the devices.
------------------------------
Can Karani Soner
------------------------------
Original Message:
Sent: 07-23-2024 04:39
From: markw
Subject: BGP Cease Error
I assume the setup is as follows:
R1<->R2 on the 10.10.10.x subnet and R1 <->R3 on the 30.30.30.x subnet.
If so, what are your subnet lengths?
Because for R1 <-> R2 I can imagine that being a /31 or something, and that makes sense for it to be 10.10.10.2 <-> 10.10.10.3
But for R1 <-> R3 I can't make sense of the subnet. If it's a /31 then 30.30.30.1 and 30.30.30.3 are not in the same subnet, and if it's a /30 then 30.30.30.3 is not a valid IP to use (it's the broadcast address), so probably leads to issues at best.
So I suspect there might be a typo in the IP addressing there?
If it's not that, could you share your interface configuration for these sessions as well for clarification?
I understand that these are fictional IPs and ASNs most likely, but just want to make sure that you're not inadvertently misconfiguring subnets here :)
Original Message:
Sent: 07-22-2024 07:51
From: Can Karani
Subject: BGP Cease Error
I received a "Code 6 Cease Error" while configuring bgp in one of my customers. And when I researched, they wrote that this error may be caused by a configuration error (MTU mismatch, Firewall Filter, Group configuration, etc.). However, there were no such mismatches in any way. I share the configurations below, they are very simple.
I did the same configurations in Juniper vlabs and I did not get this error there. I cannot share the logs exactly because my customer is a defense industry company, but the error was as follows "[code 6 (Cease) (Connection refused)"
R1; (This router has one bgp group and has two neighbor.)
group Site-Location
type external
neighbor 10.10.10.2 peer-as 65005
neighbor 30.30.30.1 peer-as 65010
R2
type external
neighbor 10.10.10.3 peer-as 65002
R3
type external
neighbor 30.30.30.3 peer-as 65002
Thats all and when ı tried this on vlabs, ı saw entire sessions are established.
Have you ever faced something like this before?
------------------------------
Can Karani Soner
------------------------------