FRRouting
tests: Update ospf6_ecmp_inter_area topotest (#16197 regression test, unified config)
The new test_multipath_inter_area() function reliably checks for #16197 and will fail until that is fixed (the current version of the topotest only fails occasionally). Also, convert topotest to use unified config.
In theory, ready for merging. But you might want to wait for the fix for #16197 to avoid permanent CI failures. Replaces #16797.
Does anybody know what this failure here means? https://ci1.netdef.org/browse/FRR-PULLREQ3-CHECKOUT-8369 The missing files really don't exist anymore, I'm using unified config (only frr.conf now)
CI failure looks related --
{
"internalNextHopActiveNum": 1,
"nexthops": [
{
"fib": true,
"interfaceName": "r1-eth2",
"active": true
}
]
}```
CI failure looks related
That's good. It shows that the new test function is working as expected since the bug it's checking for is still there. The CI failure should disappear once #16197 is fixed (by #18345, which hopefully should be ready for merging in a few days).
CI failure looks related
That's good. It shows that the new test function is working as expected since the bug it's checking for is still there. The CI failure should disappear once #16197 is fixed (by #18345, which hopefully should be ready for merging in a few days).
okay -- let me know when the other one is ready and I'll work on getting these pushed
@gromit1811 any updates on the other two pr's that will clear the errors here?
@gromit1811 any updates on the other two pr's that will clear the errors here?
@riw777 You just approved #18345, so once that is merged, this one here can go in as well. AFAIK, there's no other PR this one depends on, #16197 was just the bug report that started everything and that will be fixed by merging #18345.
The "Basic Tests" failure looks unrelated - this PR doesn't touch BGP at all. And the "Test ubuntu ARM docker image" failure looks like a stale error. The test didn't actually run after #18345 was merged - logs are from March 31.
ci:rerun ...
logs from the last run expired, need to rerun to figure out what's going on
Again, looks like nothing related to this PR. I don't see how the changes here could possibly break bgp_addpath_best_selected. And the failing ARM docker image build has been failing forever and a "ci:rerun" doesn't seem to trigger a new build. Last time I saw this elsewhere, this was due to a broken ARM Debian package mirror.
The logs for this run have expired and are no longer available. ...ugh
ci:rerun
I guess "rerun" won't help, this doesn't seem to trigger the failing build test. At least, all other tests seem to be OK now.
I could try a rebase - that should really trigger a full rebuild, shouldn't it?
ci failure doesn't look related to me ... it's over in bgp land ... I see a force push, but maybe let's try a rebase?
@Mergifyio rebase
Guess your rebase won't do much, the reason for my force push was that I did a rebase just beforehand (locally)...
The error is interesting, I've seen various BGP-related topotest failures in the past, but I can't remember that one. But it can't possibly be caused by this PR - I'm only modifying an OSPFv3 topotest here and I don't see how that could affect another topotest. And at least, the ARM build failure has finally vanished.