Hi All,
This weekend I’m going to devote a considerable amount of time to learn MPLS-TE in v7 and to me it’s seems like MT document is lack on this department, any gotcha/tips/hint or advice or is it worth an effort to implement this in MT at all?
Does the document in v6 lying around still relatively relevant on this?
Thanks in advance
examples in v6 docs are still relevant, for the most part there are only few config changes between v6 and v7 and some menus relocated.
Also use CLI.
Maris thank you, may I ask while we are at the topic what’s the equivalent of this in v7?
[admin@R1] /interface traffic-eng> monitor 0
tunnel-id: 7
primary-path-state: established
primary-path: dyn
secondary-path-state: not-necessary
active-path: dyn
active-lspid: 1
active-label: 29
explicit-route: "S:1.1.1.2/32,S:2.2.2.2/32,S:2.2.2.3/32,S:4.4.4.3/32,S:4.4.4.5/32"
recorded-route: "1.1.1.2[30],2.2.2.3[29],4.4.4.5[0]"
I can’t seem to find it could you care enough to update the documentation please?, thank you! like I said MT docs is very scarce in this area
EDIT: this command was not available in v7 or it has been moved somewhere? unsupported? how can I check if the TE tunnel has been established? again thank you!
[admin@PE2] > mpls/traffic-eng/path/print
Columns: NAME, USE-CSPF
# NAME USE-CSPF
0 dyn yes
[admin@PE2] > mpls/traffic-eng/tunnel/print
Columns: NAME, FROM-ADDRESS, TO-ADDRESS, BANDWIDTH, PRIMARY-PATH
# NAME FROM-ADDRESS TO-ADDRESS BANDWIDTH PRIMARY-PATH
0 te1 192.168.90.254 192.168.90.253 100Mbps dyn
[admin@PE2] >
[admin@PE2] > mpls/traffic-eng/flow/print #### is this supposed to be it?
[admin@PE2] >
After few days of banging my heads against the wall, yes TE is working fine with v7 you just have to work with CLI and also please improved the documentation that for the CSPF=yes and TE tunnel will establish the following condition must be met
All loopback interface must advertise in the ospf domain
All ospf router in the path should activate this in ospf instance mpls-te-address=loopback_address mpls-te-area=0.0.0.0
All Participating router (RR,P1,P2,PE1,PE2 just to give context) in the path must have participate in TE (MPLS → Traffic Eng → Interface) this is very critical the TE tunnel will just sit there doing nothing (very hard to troubleshoot)
If you don’t plan to use CSPF=yes (Dynamic) in the tunnel path then the hops parameter should be done in the CLI for now (v7.19.1) not in winbox (known issue)
All of that is already in the documentation:
https://wiki.mikrotik.com/Manual:TE_Tunnels
Like I said these v6 docs are still relevant, working principles are the same.
Yes, but the choice of words is not obvious for a non native English speaker I have to read it more than 4 times and count less of hours understanding and troubleshooting the given topology, i’ll attach here to a sample topology that can easily been understood at least from my perspective that might help others in the process, I hope MT can make a sample as easily understood as possible, anyway thanks a ton Maris for your help
@mrz can I have a quick question in absence of MPLS fast-reroute is there a way to make the convergence fast somehow, the IGP is fast with BFD but the establishment of TE tunnel is way too slow around 160seconds transition from Primary path to Secondary but from Secondary to Primary is fast around 5seconds
Is there some tunable in v6 or v7 to make it reasonable fast?
Secondary to primary is tunable by primary-retry-interval
From primary to secondary depends on how fast OSPF detects failure, routing table is updated and timeouts in TE configuration.
If CSPF is enabled you can play with reoptimize-interval, setup-priority and holding-priority in tunnel path.
Thanks I’m going to play with this parameters thanks a lot
