Hi, Here's the summary of the IRC meeting.
--- COMMUNITY MEETING Place: #openvpn-meeting on irc.freenode.net Date: Thursday 29th August 2019 Time: 20:00 CEST (18:00 UTC) Planned meeting topics for this meeting were here: <https://community.openvpn.net/openvpn/wiki/Topics-2019-08-29> Your local meeting time is easy to check from services such as <http://www.timeanddate.com/worldclock> SUMMARY cron2 and mattock participated in this meeting. --- Briefly touched on HLK testing. It looks like there is something wrong with the underlying network (or drivers) the physical HLK nodes use - there are lots of TCP-related errors during the E2EPerf test. This would help explain why the less aggressive two-machine tests pass, but E2EPerf does not. Mattock has downloaded the Virtual HLK and will convert it to Virtualbox format in the footsteps of rozmansi. With hopefully a successful e2eperf run under his belt mattock will be better equipped to figure out the issues in the physical environment. This task unfortunately progresses at a snail's pace because mattock is being stretched thin by requests from multiple groups of people. Mattock's physical HLK lab is now isolated from the rest of the office network, which will eventually (if needed) allow other people to connect to it for debugging purposes. -- Discussed the Fedora 29 connectivity issue revealed by t_client tests a couple of weeks ago. Mattock planned on separating his buildslaves, including Fedora 29, into a separate network at his home, so that access could be provided to others. However, actually placing the buildslaves to AWS EC2 would be better, as existing community VPN could then be used to provide access. Buildslave and t_client.sh setup is almost completely automated so this would not be a particularly big task. -- Noted that this Friday's mini-hackathon may have fairly limited audience. Cron2 won't make it and it is unclear if lev, dazo, ordex or arne can make it. Mattock will regardless try to move the "builder" task forward; that VM is composed of: - sbuild_wrapper: Debian/Ubuntu package builder - openvpn-build: Windows release package generation - openvpn-windows-buildtest: the "Windows buildslave" He'll also work on setting up the Virtual HLK (see above). -- Full chatlog attached. ---
20:38 < cron2> I won't make the meeting in time either - dinner will be ready around 20:05 and $wife will kill me if I prefer a meeting to "eat with family" 20:38 < cron2> mattock1: no, ordex, dazo, lev__, and plaisthos 20:38 < mattock1> $wives can be challenging :) 20:38 < mattock1> mattock3 is living inside a tmux session in irssi 20:38 < cron2> ordex sent a note at 19:25 that they want to go partying and not meet 20:38 < mattock1> are they in the same place? 20:39 < cron2> dunno, but it reads like it 20:39 < mattock1> ok, maybe they're drinking cheap beer in ukraine 20:39 < cron2> so we could have a short meeting at 20:30 or so, if you have something to report (or questions to discuss) 20:40 < mattock1> I can give that report now 20:40 < mattock1> so, having looked at packet capture logs for e2eperf test I think there's something wonky in the physical network / windows drivers 20:41 < mattock1> I downloaded the Virtual HLK and will convert that to Virtualbox format so that I can see if I can reproduce the success of rozmansi there 20:42 < mattock1> if so, I will see what a successful run should look like, then figure out how to fix the problem in the physical HLK environment 20:43 < cron2> sounds like a plan 20:43 * cron2 needs to help $wife with dinner now, will be back 20:43 < mattock1> unfortunately I'm being pulled from all directions to other tasks 20:43 < mattock1> so things progress at a snail pace 20:44 < mattock1> I started work on the Windows buildslave but it is not yet in production 20:44 < mattock1> I'll try to move that forward tomorrow in the mini-hackathon 20:45 < mattock1> plus try to convert and launch the virtual HLK thing in Virtualbox 20:56 -!- mattock2 [~mattock@openvpn/corp/admin/mattock] has joined #openvpn-meeting 21:04 < mattock1> more mattocks 21:05 < mattock1> if anyone is here let me know :) 20:38 < mattock1> are they in the same place? 20:39 < cron2> dunno, but it reads like it 20:39 < mattock1> ok, maybe they're drinking cheap beer in ukraine 20:39 < cron2> so we could have a short meeting at 20:30 or so, if you have something to report (or questions to discuss) 20:40 < mattock1> I can give that report now 20:40 < mattock1> so, having looked at packet capture logs for e2eperf test I think there's something wonky in the physical network / windows drivers 20:41 < mattock1> I downloaded the Virtual HLK and will convert that to Virtualbox format so that I can see if I can reproduce the success of rozmansi there 20:42 < mattock1> if so, I will see what a successful run should look like, then figure out how to fix the problem in the physical HLK environment 20:43 < cron2> sounds like a plan 20:43 * cron2 needs to help $wife with dinner now, will be back 20:43 < mattock1> unfortunately I'm being pulled from all directions to other tasks 20:43 < mattock1> so things progress at a snail pace 20:44 < mattock1> I started work on the Windows buildslave but it is not yet in production 20:44 < mattock1> I'll try to move that forward tomorrow in the mini-hackathon 20:45 < mattock1> plus try to convert and launch the virtual HLK thing in Virtualbox 20:56 -!- mattock2 [~mattock@openvpn/corp/admin/mattock] has joined #openvpn-meeting 21:04 < mattock1> more mattocks 21:05 < mattock1> if anyone is here let me know :) 21:05 < mattock1> it is meeting time 21:17 -!- mattock1 [~mattock@openvpn/corp/admin/mattock] has quit [Ping timeout: 244 seconds] 21:18 < mattock2> one mattock down, two to go 21:27 < cron2> now I'm back 21:28 < mattock2> hi 21:28 < cron2> this mini-hackathon will be with out me (mostly) as I'm on vacation 21:28 < cron2> I'll check in at siesta time 21:28 < mattock2> anything new at your end? 21:29 < cron2> no... the usual stuff. We need a 2.4.x release eventually, we have a few "for 2.4" patches in the queue, we want a nice and shiny tap6 driver :) and we need more active developers 21:29 < mattock2> +3 21:30 < cron2> any progress on the "debug test fail #1 on the ubuntu buildslave"? you planned to separate your network so you could give access to ordex 21:31 < mattock2> did the party animals tell if they would join the hackathon tomorrow? 21:32 < cron2> no, the only part was 21:32 < cron2> (query) 21:32 < mattock2> the HLK env is now mostly separate, but for fedora stuff creating a new buildslave in EC2 might be more useful 21:37 < mattock2> which I planned to do 21:37 -!- pippin__ [Pippin_@gateway/vpn/protonvpn/pippin/x-75792076] has joined #openvpn-meeting 21:37 < cron2> it would save a bit of effort to migrate the existing config over (t_client certs and that) 21:39 -!- mattock1 [~mattock@openvpn/corp/admin/mattock] has joined #openvpn-meeting 21:40 < mattock1> the buildslave setup is automated including t_client certs so creating Fedora 29/30 in EC2 is fairly straightforward (new keys, migration of the puppet code to OpenVPN's puppet-control-repo) 21:41 -!- Pippin_ [Pippin_@gateway/vpn/protonvpn/pippin/x-75792076] has quit [Ping timeout: 272 seconds] 21:41 < cron2> huh, so it calls out to phillip and creates keys there? nice :) 21:43 < mattock1> well that is not done at the moment, but it would be doable 21:43 < mattock1> if I recall correctly I've been shamelessly reusing vpn keys from decommissioned buildslaves 21:43 < cron2> well, that also works :) 21:45 < mattock1> anyways, the builder (sbuild_wrapper, openvpn-build, openvpn-windows-buildtest) and buildslave work is closely related - we've been moving our EC2 things under Terraform management which takes some effort but helps ongoing maintenance a lot 21:46 < mattock1> anyways, the builder VM is the first thing I'll finish (incl. windows buildslave) 21:46 < mattock1> it is "nearly there" 21:47 < mattock1> fedora 30 buildslave (which "hopefully" has the bug exposed by t_client tests) would be next 21:47 < mattock1> I hope to get builder going tomorrow - that may be a reasonable goal 21:47 < cron2> nice 21:48 < mattock1> anything else we could/should discuss? 21:48 -!- pippin__ is now known as Pippin_ 21:49 < mattock1> cron2: will you make it to tomorrow's hackathon? 21:49 < cron2> told mattock2 above 21:49 < cron2> 20:28 < cron2> this mini-hackathon will be with out me (mostly) as I'm on vacation 21:51 < mattock1> oh yes, I see that 21:51 < mattock1> so it will be me fighting openvpn-windows-buildtest 21:51 < mattock1> :P 21:54 < mattock1> so meeting concluded? 21:55 < cron2> yes :) 21:55 < cron2> 5 minutes early! 21:55 < mattock1> good, I have have to wake up very early tomorrow morning 21:55 < cron2> (has anyone seen syzzer recently?) 21:55 < cron2> good night :) 21:55 * cron2 won't get up before 8:30... kids might get up at 9:30 21:55 < mattock1> no, I have not 21:55 < mattock1> lucky you :) 21:56 < mattock1> I have to get up in 9 hours and I'm only eating a bit before heading to shower and bed 21:56 < mattock1> anyways good meeting even with the limited participants we had :D 21:58 < mattock1> summary tomorrow
signature.asc
Description: OpenPGP digital signature
_______________________________________________ Openvpn-devel mailing list Openvpn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/openvpn-devel