On Wed, 20 Apr 2016, David Lamparter wrote:
The link-detect comparison keeps coming up, yet there's a major
difference: When I argued for the "slow" link-detect migration, I was
personally aware of 2 installations that would've been broken by the
change, one of them /my own network/.
And that's fine.
I hate it when some software I use goes and changes behaviour. I like
stability in the software I use, and I like having control over change.
So, if a patch is proposed, and others note "Hmm, that could break XYZ"
and if that leads to patches with better transition paths, that's great.
It's hard to measure the value, and the costs are definite, but we've been
conservative in the past. And overall, it's probably better to be
conservative.
Paul (or anyone else), is there an actual user installation you're aware
of that would be broken by this change? If yes, could you please
describe it?
See my other replies. The first patch, by just flipping the default, just
creates the problem in the interim on networks that used to work (Quagga
networks). And there was no config option either.
If it's a serious problem, the first patch needed a bit more work (and my
patch-set maybe should build on it actually, or at least ACK it).
If it's not a serious problem, then we still need the config options, and
we've definitely got time to see if H-bit is serious and if so, how to
best transition over to that with the minimum of behavioural churn.
Surely?
regards,
--
Paul Jakma, HPE Aruba, Advanced Technology Group
Fortune:
Quantum Mechanics is a lovely introduction to Hilbert Spaces!
-- Overheard at last year's Archimedeans' Garden Party
_______________________________________________
Quagga-dev mailing list
[email protected]
https://lists.quagga.net/mailman/listinfo/quagga-dev
