frr
frr copied to clipboard
BMP Adj-RIB-Out (RFC8671), ECMP Support and Loc-RIB missing features
Global:
- added peer type for all monitored ribs
- added peer distinguisher for all monitored ribs
- include mpls labels (when possible, no adj-in-pre)
Add support for Adj-RIB-Out monitoring in BMP:
- pre- and post-policy
- config knobs
- sync
- incremental updates
- stats (when possible)
- no timestamp (currently not recorded in adj-out)
Add ECMP support in BMP:
- include addpath id in bmp in adj-in pre/post, loc-rib and adj-out pre/post
- monitoring on all ecmp selected path (BGP_PATH_SELECTED | BGP_PATH_MULTIPATH)
Add Loc-RIB missing features:
- peer up/down on vrf state update
- missing stats
Missing:
- ?
ci:rerun
This pull request has conflicts, please resolve those before we can evaluate the pull request.
@ton31337 would you prefer if i squash them together?
This pull request has conflicts, please resolve those before we can evaluate the pull request.
@ton31337 would you prefer if i squash them together?
Yes, please do ...
@ton31337 @riw777 done. happy new year btw
Also2: I don't get it why https://github.com/FRRouting/frr/commit/fa98a15ce02814d7313eba11fef84cdaab76fff3 commit is a separate from the main commit where the main locking/unlocking for path info happens?
This commit fixes (thanks to @donaldsharp) an issue that only happens when the bmp module is loaded. I think it is interesting to keep it separate because it shows that the code looks like it does because of this issue.
This pull request has conflicts, please resolve those before we can evaluate the pull request.
This pull request has conflicts, please resolve those before we can evaluate the pull request.
@mxyns still working on this?
@riw777 yes. i think i addressed every comment i had at this point, no? i just keep fixing merge conflicts
CI:rerun Rerun after fixing git access on CI infra
there were lots of conflicts let's hope i didn't mess something up
ci:rerun
looks like some related ci failures yet ...
This pull request has conflicts, please resolve those before we can evaluate the pull request.
looks like some related ci failures yet ...
i only see convergence and unrelated evpn issues, isn't it?
now i have conflics again though...
@riw777 ci still not passing but the failing tests look to me like they are not related to the changes, any opinion?
looks good, just need to sort topo
Great thanks. Any idea how ? Most of these are convergence time issues..
This pull request has conflicts, please resolve those before we can evaluate the pull request.
I guess we need to fix the conflicts now before we can get back to the failures ... rebasing might help with the ci failures anyway
This pull request has conflicts, please resolve those before we can evaluate the pull request.