I am not sure I follow the topology correctly, but I think you might be breaking some basic OSPF area rules.
There can only be one of each area and the area must be fully connected for all devices in the area. You cannot have two area 0 or two area 1 or two area 2.
Your diagram seems to imply that there are two area 0 that have no connection to each other and operate independantly.
Hello Juniper Forum,
As far as I can tell, I have all of the components set for MVPN on both PE's.
I am wondering if there is a trigger to bring up teh MVPN neighbor (provider tunnel). This seems like the first step that needs to happen -- in the order of troubleshooing MVPN connectivity.
What is missing? PE configs attached in above post.
======
SomeEngineer@PE-1# run show mvpn neighbor
MVPN instance:
<snip legend>
Family : INET
Instance : CPE-1
MVPN Mode : RPT-SPT
MVPN instance:
<snip legend>
Family : INET6
Instance : CPE-1
MVPN Mode : RPT-SPT
SomeEngineer@PE-1# run show mvpn instance
MVPN instance:
<snip legend>
Family : INET
Instance : CPE-1
MVPN Mode : RPT-SPT
Provider tunnel: I-P-tnl:LDP-P2MP:1.1.1.1, lsp-id 16777217
MVPN instance:
<snip legend>
Family : INET6
Instance : CPE-1
MVPN Mode : RPT-SPT
Provider tunnel: I-P-tnl:LDP-P2MP:1.1.1.1, lsp-id 16777217
yes,spuluka.
thanks very much for your reply.
there are 2 seperated area 0 now.
1:I am wondering which issue I should see now(it was confgured wrong I guess).
2: the way to fix it is to create a vlan and assign ip to it, then enable it va the trunk link,right?
3:devices in area1 or area2 are fully connected, right?
Hi,
Sorry for the delay in response. Can you please share the output of "show ldp p2mp tunnel extensive" from both PEs to check why the tunnel is not getting formed? From your configuration on PE1, I see you have used 2 IP addresses on lo0 address and you have specified router-id as transport-address under ldp. Just to make sure we are not seeing any issues with that, can you please try to remove one of the addresses from lo0 on PE1 and check once?
If you see a different address being used to form the LDP session and you also want to keep 2 addresses on lo0, then you can also configure following and check:
set routing-options multicast local-address
Hope this helps.
Thanks
1:I am wondering which issue I should see now(it was confgured wrong I guess).
Your OSPF route tables will not be complete and some devices won't be able to reach each other as a result.
2: the way to fix it is to create a vlan and assign ip to it, then enable it va the trunk link,right?
Yes, if you add a path and connect the two area 0 then this would work.
3:devices in area1 or area2 are fully connected, right?
I think they are because you say the trunk port betweeen the two sides carry those vlans and I assume those interfaces are placed into the correct area.
Thanks for your help with this.
"show ldp p2mp tunnel extensive" #This command did not work.
SomeLabGuy@PE-1> show ldp ?
Possible completions:
database Show LDP label database
fec-filters Show filters for forwarding equivalence classes
interface Show LDP interface status
neighbor Show LDP neighbor status
oam Show LDP OAM information
overview Show LDP overview information
path Show LDP path table
route Show LDP internal topology table
session Show LDP session status
statistics Show LDP statistics
traffic-statistics Show packet statistics for LDP label-switched paths
PE-1: Lo1 address changed
set interfaces lo0 unit 0 family inet address 1.1.1.1/32
set interfaces lo0 unit 1 family inet address 11.11.11.11/32
set routing-instances CPE-1 interface lo0.1
PE-2: Lo2 address changed
set interfaces lo0 unit 0 family inet address 4.4.4.4/32
set interfaces lo0 unit 2 family inet address 44.44.44.44/32
set routing-instances CPE-2 interface lo0.2
NOTE: since I keep playing with this lab, I am attaching the current configs.
L3VPN still up/working MVPN - PE-provider tunnels still not up.
CPE-1(lo0:100.100.100.100)
PE1(1.1.1.1)----P1(2.2.2.2)----P2(3.3.3.3)----PE2(4.4.4.4)
(lo0:200.200.200.200)CPE-2
SomeLabGuy@CPE-1> traceroute 200.200.200.200 source 100.100.100.100
1 192.168.1.1 (192.168.1.1) 2.541 ms 2.916 ms 2.468 ms<--PE-1
2 12.12.12.2 (12.12.12.2) 3.547 ms 3.669 ms 3.217 ms <--P1
MPLS Label=299808 CoS=0 TTL=1 S=0
MPLS Label=16 CoS=0 TTL=1 S=1
3 23.23.23.3 (23.23.23.3) 3.761 ms 3.937 ms 3.628 ms <--P2
MPLS Label=299808 CoS=0 TTL=1 S=0
MPLS Label=16 CoS=0 TTL=2 S=1
4 172.16.2.200 (172.16.2.200) 4.188 ms * 4.985 ms <-CPE-2
SomeLabGuy@PE-1> show ldp neighbor brief
Address Interface Label space ID Hold time
12.12.12.2 fe-0/0/7.0 2.2.2.2:0 11
SomeLabGuy@PE-2> show ldp neighbor
Address Interface Label space ID Hold time
34.34.34.3 fe-0/0/7.0 3.3.3.3:0 11
router1 has 1 link to switch1 and another switch2, only one link is active at the same time(it switch to other link sometimes).
1: should I add router1 into ospf area3 and advertise the 192.168.1.0/24?
or redistribute it to ospf in switch1 and switch2?
what is the difference
Hello ,
I have to configure OTN Interface with FPC MPC5E 3D 24XGE+6XLGE on MX960.
Please help me , How to configure the OTN on mx 960?
Thanks & Regards,
Does vMX support sFlow v9? I cannot find it in patfinder. Thanks.
HI, yes jflow v9 should work on vMX
Hello,
I am assuming the blue router and the two green ones are all different.
If only one of the links (r1-s1 and r1-s2) is active at any given time then it would be easier to configure the blue router in ospf and redistribute from there.
Here are some disadvantages from redistributing the route from s1 and s2:
- You would have to depend on static routing.
- If for any reason the active interface (r1-s1 or r1-s2) looses connectivity but remains UP, then the static route would remain active and traffic would be blackholed.
If you add the blue router in ospf and redistribute from there:
- OSPF will take care of the routing when the connection changes from switch1 to switch2.
- OSPF hello's would also serve as a mechanism to detect failures in the link (automatic failover).
I hope this helps.
Hello,
Does anyone have a recommendation to move forward with this lab testing.
Are there other Juniper forums where I can get some help with this lab? MVPN is not a very common topic.
Problem solved:
Removed these lines from the PE configs. auto-discovery found and brought up the MVPN neighbors.
NOTE: Still testing....to verify multicast is still working between the CE's. currently my RP announce/discovery messaging is sync'ing between the CE's. Will now test that joins/prunes work correctly.
PE-1
delete protocols mpls label-switched-path R1-to-R4 to 4.4.4.4
PE-2
delete protocols mpls label-switched-path R4-to-R1 to 1.1.1.1
PE-1# run show mvpn neighbor
MVPN instance:
<snip legend>
Family : INET
Instance : CPE-1
MVPN Mode : RPT-SPT
Neighbor I-P-tnl
4.4.4.4 LDP-P2MP:4.4.4.4, lsp-id 16777218
PE-1# run show mvpn instance
MVPN instance:
<snip legend>
Family : INET
Instance : CPE-1
MVPN Mode : RPT-SPT
Provider tunnel: I-P-tnl:LDP-P2MP:1.1.1.1, lsp-id 16777217
Neighbor I-P-tnl
4.4.4.4 LDP-P2MP:4.4.4.4, lsp-id 16777218
C-mcast IPv4 (S:G) Ptnl St
100.100.100.100/32:224.0.1.39/32 LDP-P2MP:1.1.1.1, lsp-id 16777217 RM
200.200.200.200/32:224.0.1.40/32 LDP-P2MP:4.4.4.4, lsp-id 16777218
0.0.0.0/0:225.1.1.1/32 LDP-P2MP:1.1.1.1, lsp-id 16777217
0.0.0.0/0:239.2.2.2/32 LDP-P2MP:1.1.1.1, lsp-id 16777217 RM
100.100.100.100/32:239.2.2.2/32 LDP-P2MP:1.1.1.1, lsp-id 16777217 RM
PE-2# run show mvpn neighbor
MVPN instance:
<snip legend>
Family : INET
Instance : CPE-2
MVPN Mode : RPT-SPT
Neighbor I-P-tnl
1.1.1.1 LDP-P2MP:1.1.1.1, lsp-id 16777217
PE-2# run show mvpn instance
MVPN instance:
<snip legend>
Family : INET
Instance : CPE-2
MVPN Mode : RPT-SPT
Provider tunnel: I-P-tnl:LDP-P2MP:4.4.4.4, lsp-id 16777218
Neighbor I-P-tnl
1.1.1.1 LDP-P2MP:1.1.1.1, lsp-id 16777217
C-mcast IPv4 (S:G) Ptnl St
100.100.100.100/32:224.0.1.39/32 LDP-P2MP:1.1.1.1, lsp-id 16777217
200.200.200.200/32:224.0.1.40/32 LDP-P2MP:4.4.4.4, lsp-id 16777218 RM
0.0.0.0/0:239.2.2.2/32 LDP-P2MP:1.1.1.1, lsp-id 16777217
100.100.100.100/32:239.2.2.2/32 LDP-P2MP:1.1.1.1, lsp-id 16777217
Good day!
I have a policy on Cisco ASR:
interface Bundle-Ether1.100 service-policy input PM-SET-DSCP-DEFAULT service-policy output PM-SET-DSCP-DEFAULT ....
policy-map PM-SET-DSCP-DEFAULT
class class-default
set dscp 0
!
end-policy-map
! So i set dscp 0 on ingress and egress traffic.
I need to do the same thing on Juniper MX480
I have made this:
class-of-service {
interfaces {
ae1 {
unit 100 {
rewrite-rules {
dscp PM-SET-DSCP-DEFAULT-OUT;
}
}
}
}
rewrite-rules {
dscp PM-SET-DSCP-DEFAULT-OUT {
forwarding-class network-control {
loss-priority low code-point 000000;
loss-priority high code-point 000000;
}
forwarding-class assured-forwarding {
loss-priority low code-point 000000;
loss-priority high code-point 000000;
}
forwarding-class best-effort {
loss-priority low code-point 000000;
loss-priority high code-point 000000;
}
forwarding-class expedited-forwarding {
loss-priority low code-point 000000;
loss-priority high code-point 000000;
}
}
}But i still can't find how to set dscp on ingress packets. Could anybody help?
Hi,
An MF classifier configured as input firewall filter on the interface should be able to mark and classify selected packets on ingress:
Alternatively you could use a fixed classifier on the interface to classify all ingress packets and rewrite that forwarding class with the desired dscp marking on egress.
Cheers,
Ashvin
Hi,
We have recently upgraded our MX480 router from JUNOS 13.3R6.5 built to the JUNOS 15.1F6.9 along with JSDN package support. After the upgrade we are seeing some different values in the show system processes extensive O/P. The output is given below:
show system processes extensive
last pid: 32251; load averages: 0.26, 0.27, 0.25 up 4+10:46:54 12:47:52
128 processes: 2 running, 125 sleeping, 1 waiting
Mem: 243M Active, 1669M Inact, 400M Wired, 272K Cache, 92M Buf, 916M Free
Swap: 8192M Total, 32K Used, 8192M Free
PID USERNAME THR PRI NICE SIZE RES STATE C TIME WCPU COMMAND
10 root 4 155 ki31 0K 48K CPU3 3 416.0H 399.56% idle
22180 root 2 -26 r26 892M 30064K nanslp 1 30:12 3.76% chassisd
My question is that why it is showing the WCPU value in 399% form, why it is not showing the values like others in the range from 0-100.
Thanks
As shown in the diagram above, there are area 0 ,2 and 5 in ospf. it is running well now
since 3.3.3.0/24 and 4.4.4.0/24 has heavy traffic ,we plan to add 2 links to dedicated for these prefix (2 red links)
how to achieve in my case ?
That is because your RE is re-1800x4 . This means that it is 4 core cpu. and utilization is shown 100% for each core.
while you running on 13.3 juns use only single core of 4 available and show you cpu usage as 100% maximum
Hi all,
In the context of L3VPN (family inet-vpn unicast), what's the difference between:
set protocols bgp group <group> multipath
and:
set routing-instances <vrf> routing-options multipath
If I have the same prefix announced from two PEs, enabling multipath at the BGP level works just fine. So what does multipath at the routing-instances level give me extra? When should I use one versus the other? Or should I enable both?
Thanks,
Bob
