bfd-topo1: add basic topology test for bfdd
#96
Conversation
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
|
@rzalamena can you rebase when you have the chance |
bfd-topo1/test_bfd_topo1.py
Outdated
| if tgen.routers_have_failure(): | ||
| pytest.skip(tgen.errors) | ||
|
|
||
| topotest.sleep(2, 'waiting for bfd peers to go up') |
There was a problem hiding this comment.
I'm not a fan of sleep dependent tests -- sleep interval may vary in different test environments or even run to run in the same environment. Better to pool for a result with a max allowed time.
There was a problem hiding this comment.
This should not be a problem the BFD protocol converges in less than a second, that was added as a safety mesure but probably not needed. If it takes more than 3 seconds then bfdd already failed.
It seems that recording the time taken for a function is a repeating pattern that we should think about implementing in topotests/topogen.
bfd-topo1/test_bfd_topo1.py
Outdated
| if tgen.routers_have_failure(): | ||
| pytest.skip(tgen.errors) | ||
|
|
||
| topotest.sleep(5, 'waiting for bgp peers to go up') |
There was a problem hiding this comment.
Same comment as above, why is this sleep needed given that run_and_expect is used?
bfd-topo1/test_bfd_topo1.py
Outdated
|
|
||
| # Wait the minimum time we can before checking that BGP/BFD | ||
| # converged. | ||
| topotest.sleep(1, 'waiting for BFD to converge') |
I rebased the branch, but I'll fix the sleep as next thing. I'm thinking about something to record the time taken to converge and I'll come back later. Thanks for the review! |
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
Continuous Integration Result: SUCCESSFULCongratulations, this patch passed basic tests Tested-by: NetDEF / OpenSourceRouting.org CI System CI System Testrun URL: https://ci1.netdef.org/browse/FRR-TOPOPR-328/ This is a comment from an EXPERIMENTAL automated CI system. |
Continuous Integration Result: SUCCESSFULCongratulations, this patch passed basic tests Tested-by: NetDEF / OpenSourceRouting.org CI System CI System Testrun URL: https://ci1.netdef.org/browse/FRR-TOPOPR-329/ This is a comment from an EXPERIMENTAL automated CI system. |
|
@louberger is this fixed to your satisfaction now? |
Allow BFDd to be configured and used. Signed-off-by: Rafael Zalamena <[email protected]>
Test if BFD peers have found each other and if the BGP neighors have connected. Signed-off-by: Rafael Zalamena <[email protected]>
Test that after a link goes down BGPd will be notified and recovered quickly. Also test that BFD show command tells us that the peer went down. Signed-off-by: Rafael Zalamena <[email protected]>
Import the graphviz file and the generated picture. Signed-off-by: Rafael Zalamena <[email protected]>
|
Rebased with |
Continuous Integration Result: SUCCESSFULCongratulations, this patch passed basic tests Tested-by: NetDEF / OpenSourceRouting.org CI System CI System Testrun URL: https://ci1.netdef.org/browse/FRR-TOPOPR-356/ This is a comment from an EXPERIMENTAL automated CI system. |
This PR implements a topology test for the new
bfdddaemon ( FRRouting/frr#2294 ).The following tests are done: