I agree with Bryan. I have been testing in production without H2 (waiting for an internal plugin to upgrade). With three patches I am able to run for a couple hours without crashing. That crash is reported as issues 1480 and 1476. Sadly due to delays in getting our internal plugins upgraded, I have only been able to seriously work the 7.x build that last couple weeks. But 7.x is the top priority for our team right now
Sent from Yahoo Mail on Android On Mon, 27 Feb 2017 at 6:36 pm, Bryan Call<bc...@apache.org> wrote: Right now I would pick number 2. I have been able to get 7.1.x to run for awhile with a limited amount of traffic (~550 rps). I am having problems with our Carp plugin that is making it difficult for me to throw a lot of traffic at ATS 7.1.x. -Bryan > On Feb 27, 2017, at 2:28 PM, Leif Hedstrom <zw...@apache.org> wrote: > > Hi all, > > tldr; I’m struggling getting v7.1.0 into a usable state :-/. I’m to the point > now where I’m thinking we either cancel 7.1.0 entirely, unless something > changes, or we have to work together on fixing things. Below are the current > show stopping issues, me and my team are working on the first 3 of these > (since we have these problems in production), but the other open issues have > had no traction. I tried hard to isolate one of them, which was introduced > late in the cycle, but due to the nature of how rarely it triggers (once > every 1-3 days), I could not track it down. > > So, here are some thoughts / options: > > 1) We cancel v7.1.0 entirely, and focus on making 7.2.0 stable. > > 2) More people / companies help out on fixing the 7.0.0/7.1.0 open > issues, asap. > > 3) We change RM for 7.1.0, in the hope that she can make 7.1.0 stable / > usable. > > > I’ll also request a 30min slot for the upcoming Summit, to discuss release > processes. At this point, seeing how slow we are at using / upgrading > versions of ATS in production, I think can only justify one release / year, > with bug fixes in between. Basically, basically every release would be LTS, > and likely for a longer time period (2-3 years). > > Thanks, > > — leif > > 1427 - Crash in ATSConsistentHash::lookup_by_hashval > <https://github.com/apache/trafficserver/issues/1427> > 1335 - Deadlock in HostDB > <https://github.com/apache/trafficserver/issues/1335> > > > 1326 - Core dump in SSLAccept / SSLNetVConnection::sslServerHandShakeEvent > <https://github.com/apache/trafficserver/issues/1326> > 1412 - Assertion in HttpSM::tunnel_handler_post > <https://github.com/apache/trafficserver/issues/1412> > 1401 - Segfault in write_to_net_io with 7.1.x > <https://github.com/apache/trafficserver/issues/1401> > 1421 - Segmentation fault on TLS when destination server issues Reset > <https://github.com/apache/trafficserver/issues/1421> > 1363 - crash in HostDBContinuation::dnsEvent with ATS 7.0.0 > <https://github.com/apache/trafficserver/issues/1363> > 1332 - Possibly crasher in HostDB serialization (disk sync) > <https://github.com/apache/trafficserver/issues/1322> > 1480 - Crash in Http1ClientSession::set_inactivity_timeout > <https://github.com/apache/trafficserver/issues/1480> > 1476 - Crash in get_client_addr() > <https://github.com/apache/trafficserver/issues/1476> >
