Good catch Rob and I think most of us understand that this is a big change that moves ATC forward and in the right direction so it's understandable that there will be some issues found.
Jeremy On Fri, Jan 3, 2020 at 5:49 PM Robert O Butts <[email protected]> wrote: > -1 Found a serious issue in the config gen: > https://github.com/apache/trafficcontrol/pull/4256 > > TLDR The meta config includes nonexistent DSes for old location params (and > always has), but because the Config Generator uses the API, it fails to get > URL Sig endpoints for DSes that don't exist. Because it's using the API, > there's no reasonable way to make it successfully return keys for DSes that > don't exist anymore. The PR fixes the meta to omit DSes that don't exist, > and log a warning. > > This is a serious issue, and has a very high likelihood of breaking > production systems. It's caused by bad data, but the bad data is very > likely to exist for most users. > > Sorry. And apologizes in advance for more of these. Like I keep saying, the > config gen is a massive amount of code rewrite, I've done and am still > doing a great deal of parity testing, but 20k lines is more than anyone can > write without bugs. It's pretty likely there'll be a few more of these. :( > > Thanks @traeak for finding this issue. And thanks everyone, for being > patient with this large refactor. > > > On Wed, Dec 18, 2019 at 12:59 PM Dave Neuman <[email protected]> wrote: > > > I second your sentiment, Rawlin. > > > > On Wed, Dec 18, 2019 at 10:40 AM Rawlin Peters <[email protected]> > wrote: > > > > > In the spirit of getting releases out every 4-6 weeks, I'd like us to > > > refrain from downvoting the RC for minor issues. If we cut new RCs for > > > every minor issue that comes up, we'll never reach our goal of a 4-6 > > > week release period. That said, I'd like to see downvotes for major > > > regressions or major bugs that would absolutely prevent us from taking > > > this release to a production environment. If the issue doesn't fall > > > into that category, assume it will be fixed in the following release > > > in 4-6 weeks. > > > > > > - Rawlin > > > > > > On Wed, Dec 18, 2019 at 8:13 AM ocket 8888 <[email protected]> > wrote: > > > > > > > > I gotta be -1 because I broke the docs - and that only takes a second > > to > > > > fix. There's a busted link on the page for `/staticdnsentries`. > > > > > > > > I opened a PR to fix it in master: > > > > https://github.com/apache/trafficcontrol/pull/4220 which I can also > > > submit > > > > against 4.0.x in a second > > > > > > > > On Tue, Dec 17, 2019 at 12:51 PM Rawlin Peters <[email protected]> > > > wrote: > > > > > > > > > Hello All, > > > > > > > > > > I've prepared a release for v4.0.0-RC0 > > > > > > > > > > The vote is open for at least 72 hours and passes if a majority of > at > > > > > least 3 +1 PMC votes are cast. > > > > > > > > > > [ ] +1 Approve the release > > > > > > > > > > [ ] -1 Do not release this package because ... > > > > > > > > > > Changes since 3.1.0: > > > > > > > > > > > > > > > > https://github.com/apache/trafficcontrol/compare/RELEASE-3.1.0...RELEASE-4.0.0-RC0 > > > > > > > > > > This corresponds to git: > > > > > Hash: aff0186d5c982c00e82d8aad36e2a6c299049f5f > > > > > Tag: RELEASE-4.0.0-RC0 > > > > > > > > > > Which can be verified with the following: git tag -v > > RELEASE-4.0.0-RC0 > > > > > > > > > > My code signing key is available here: > > > > > > http://keys.gnupg.net/pks/lookup?search=0x8A0712500C70C06E&op=vindex > > > > > > > > > > Make sure you refresh from a key server to get all relevant > > signatures. > > > > > > > > > > The source .tgz file, pgp signature (.asc signed with my key from > > > > > above), and sha512 checksums are provided here: > > > > > > > > > > https://dist.apache.org/repos/dist/dev/trafficcontrol/4.0.0/RC0 > > > > > > > > > > > > > > > Thanks! > > > > > Rawlin Peters [email protected] > > > > > > > > > > >
