16:01:05 #startmeeting tor anti-censorship meeting 16:01:05 here is our meeting pad: https://pad.riseup.net/p/r.9574e996bb9c0266213d38b91b56c469 16:01:05 editable link available on request 16:01:05 Meeting started Thu Mar 20 16:01:05 2025 UTC. The chair is shelikhoo. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:01:05 Useful Commands: #action #agreed #help #info #idea #link #topic. 16:01:21 hello 16:01:24 hi~hi~ 16:03:07 hi 16:05:09 okay let's start the meeting 16:05:23 I didn't see a lot of new topics 16:05:28 the first topic is from me 16:05:34 Snowflake Staging Server & rdsys containerization 16:06:00 the context was that both rdsys and snowflake staging server is considering using container to automate the deployment 16:06:14 https://gitlab.torproject.org/tpo/tpa/team/-/issues/41769 16:06:32 https://gitlab.torproject.org/tpo/tpa/team/-/issues/42080 16:06:53 and we are facing some constrains with existing systems 16:07:17 the most notable one is that from each branch, it is not possible to create more than one matching host name 16:07:54 This does not work really well with snowflake as both server and broker will need their own domain name 16:08:14 otherwise, there will be an need of additional router in front of it 16:08:19 do they have conflicting API paths? could we expose both in the same domain name? 16:08:30 in theory we could 16:08:39 but a custom router will be needed 16:09:08 at least only the server accepts websocket connection 16:09:20 my current plan is try to not reinvent any wheel 16:09:31 sure 16:10:03 by setup a rootless-kubernetes + acme hosting system 16:10:12 and interact with it using terraform 16:10:28 so deployments can be automated 16:10:41 with standardized tools 16:10:51 that sounds like a lot of work to setup and maintain, but I have no experience with kubernetes 16:11:04 and currently the rootless-kubernetes + acme hosting system is done 16:11:28 ohh, you already setup kubernetes? 16:12:06 I was trying to get terraform to work with deploying snowflake 16:12:08 yes it took maybe 2 days of experiment 16:12:13 but it is already running now without root 16:12:29 with acme to renew wildcard certificate 16:13:10 and currently experimenting with terraform files to deploy snowflake servers 16:13:44 my question was is this a setup we are looking forward to 16:14:38 or I should maybe try with something like podman compose, which is not really standardised 16:14:48 from one side I'm worried no one besides you might be able to fix it 16:15:01 and I think everybody in the team knows docker/podman compose 16:15:30 I know TPA wants to support kubernetes in the future, so this setup might be migratable to their kubernets once it exists 16:15:33 how difficult is it to learn how to use? 16:15:34 but that might take years 16:15:59 like if you set everything up, will it take a long time for us to learn how to deploy different versions of snowflake/rdsys? 16:16:17 I think once there is a manual, it is not hard to get started 16:16:36 with terraform it should just take few min to learn 16:16:37 but = 16:16:48 if there is a need to change certain configuration 16:17:12 like adding a new component 16:17:24 then it will take a few days to learning, I think 16:17:32 (rdsys can live with one domain for staging, I was planning to use docker-compose and the TPA system for it) 16:17:39 yes 16:18:47 i'm not opposed to trying out this system you built for snowflake shelikhoo 16:18:56 so we are not objecting it, right now, but more information is needed 16:19:09 we wouldn't have the change the rdsys staging deployment right? 16:19:35 I think this is specific for snwoflake 16:19:36 no, snowflake and rdsys don't have to use the same system 16:19:54 but might make sense to have homogenety on how we do staging 16:20:02 thanks, cohosh. 16:20:14 let's try your system out 16:20:23 in that case, it could be interesting to try it out then, i mostly don't want it to be a big maintenance cost considering we're outside of TPA's scope 16:20:37 but it could be fun to try something different and compare the setups 16:20:45 I will keep experimenting with it and once it is ready for showcase, I will let everyone know 16:20:46 yes 16:20:57 nice 16:21:07 I have designed it to be very automate-able 16:21:29 yes, I think this topic is finished 16:21:56 now there is a interesting link: 16:21:59 https://github.com/doudoulong/Minecruft-PT/blob/main/README.md#tor-browser-traffic-tunneling 16:22:32 anything discussion about this minecraft pt? 16:23:35 not from me 16:23:41 okay 16:24:05 anything we would like to discuss today? 16:24:12 anything more we would like to discuss today? 16:24:30 does anyone want to do a reading group for two weeks from now? 16:24:52 sure 16:25:04 yes 16:25:07 any proposals? 16:25:37 I recal this paper from a previous interesting links: https://arxiv.org/abs/2409.06247 16:25:55 yeah i've read that one, it's interesting 16:25:58 i'd like to discuss it 16:26:22 then let's do that one in two weeks 16:26:55 April 3 16:27:14 yes! 16:27:56 anything more we would like to discuss? 16:28:03 not from me 16:28:25 #endmeeting