=== mup_ is now known as mup | ||
cjwatson | Any objections to me self-approving https://code.launchpad.net/~cjwatson/lazr.restful/py3-print/+merge/378511 ? I don't think it makes sense to have somebody else spend their time reading a bulk print syntax conversion | 11:56 |
---|---|---|
ilasc | cjwatson: no objections from my side | 11:58 |
cjwatson | landed, thanks | 12:16 |
cjwatson | https://code.launchpad.net/~cjwatson/lazr.restful/wsgiref-py2/+merge/378432 wants an actual review, I think, but is also very short | 12:17 |
cjwatson | And https://code.launchpad.net/~cjwatson/lazr.restful/py3-raise/+merge/378513 | 12:23 |
cjwatson | I'd also like reviews of https://code.launchpad.net/~cjwatson/launchpad/+git/launchpad/+merge/378430 and https://code.launchpad.net/~cjwatson/launchpad/+git/launchpad/+merge/378459, please - more various dependency mangling for py3 | 12:47 |
cjwatson | https://code.launchpad.net/~cjwatson/lazr.restful/six-urllib/+merge/378515 | 13:19 |
cjwatson | And https://code.launchpad.net/~cjwatson/launchpad/+git/launchpad/+merge/378518 to upgrade pip | 13:26 |
tomwardill | cjwatson: GIT_PROXY_COMMAND appears to not be doing anything | 14:20 |
tomwardill | I've edited the script to create a file in /tmp when it's run | 14:21 |
tomwardill | but no file appears | 14:21 |
tomwardill | "the script" = snap-git-proxy | 14:21 |
cjwatson | tomwardill: Which git invocation is failing? | 14:23 |
tomwardill | clone | 14:23 |
cjwatson | tomwardill: Outside the docker container? | 14:23 |
tomwardill | yep | 14:23 |
tomwardill | https://pastebin.canonical.com/p/vWjcSTgWR9/ | 14:23 |
cjwatson | tomwardill: Have you tried stracing git? | 14:23 |
tomwardill | not yet, my strace parsing is not good. | 14:24 |
tomwardill | will give it a try though | 14:24 |
cjwatson | add strace to deps and wrap strace -f -s1024 -tt around the git call | 14:24 |
cjwatson | then you should get something useful in the build log | 14:24 |
cjwatson | happy to have a look / walk you through it if you can get a trace | 14:25 |
tomwardill | righto, trying | 14:25 |
cjwatson | but you could start by searching the trace for snap-git-proxy and seeing if it's e.g. trying to exec it but failing | 14:25 |
cjwatson | Can I have a review of https://code.launchpad.net/~cjwatson/launchpadlib/py38-access-token/+merge/378523 to fix a crash on focal please? | 14:30 |
tomwardill | sometimes, I dislike pythons string syntax if you forget the , in a list | 14:30 |
tomwardill | looking | 14:31 |
tomwardill | cjwatson: +1 | 14:32 |
cjwatson | Thanks | 14:35 |
tomwardill | cjwatson: snap-git-proxy doesn't appear in the strace: https://pastebin.canonical.com/p/Dv4vNfh5b2/ | 14:41 |
tomwardill | (I ctrl-c'd when it looked like it had entered a polling/timeout loop) | 14:41 |
cjwatson | tomwardill: Could you add -v to the strace arguments so that we get unabbreviated versions of the environment variables in subprocesses? | 14:51 |
cjwatson | tomwardill: GIT_PROXY_COMMAND is actually only for git:// | 14:51 |
tomwardill | aha! | 14:52 |
cjwatson | tomwardill: For https://, git should honour https_proxy | 14:52 |
tomwardill | aha! aha! | 14:52 |
tomwardill | right. | 14:52 |
tomwardill | that might explain it | 14:52 |
tomwardill | I'll get a trace with -v and then try and work out why https_proxy isn't behaving | 14:52 |
tomwardill | cjwatson: okay, it looks like what ever is meant to be running on the buildd machine on 8222 isn't actually running | 15:04 |
tomwardill | looks like that's started in startProxy in SnapBuildProxyMixin | 15:05 |
tomwardill | oorrrr... there' | 15:06 |
tomwardill | there's a firewall problem | 15:06 |
cjwatson | Did you remember to call that? | 15:06 |
cjwatson | The local proxy on 8222 should write access logs to the build log | 15:07 |
cjwatson | lpbuildd/oci.py: args.extend(self.startProxy()) | 15:07 |
cjwatson | Hm looks OK | 15:07 |
tomwardill | yeah, i mostly suspect firewall | 15:07 |
tomwardill | just constructing a test environment to find out | 15:07 |
cjwatson | OK, if you don't get anywhere that way, I'd consider stracing (only) the top-level launchpad-buildd process as the build is running and looking for CONNECT requests | 15:08 |
tomwardill | aha! progress | 15:24 |
tomwardill | `ufw allow in on lpbuilddbr0 to any port 8222` did the trick | 15:25 |
tomwardill | and now I'm failing in docker, which is more what I expected | 15:25 |
cjwatson | Ah good | 15:25 |
tomwardill | https://docs.docker.com/config/daemon/systemd/ well, that proxy section is a bit ergh | 16:07 |
SpecialK|Canon | specifically...? | 16:11 |
cjwatson | Easier if you put stuff in place before installing docker, I expect (since then you don't need to restart anything) | 16:17 |
tomwardill | one step further down the line... | 16:39 |
tomwardill | 2020-02-04 16:56:20+0000 [-] Returning build status: OK | 16:56 |
tomwardill | IT LIVES! | 16:56 |
SpecialK|Canon | w00t | 16:56 |
tomwardill | okay, wasn't actually all that bad a change, just took a while to get there... | 16:57 |
SpecialK|Canon | tomwardill: this Issue just went past on my timeline and I thought of you https://github.com/docker/for-linux/issues/690 | 17:48 |
tomwardill | SpecialK|Canon: that’s the same problem as the lxd one I mentioned in the meeting | 17:59 |
SpecialK|Canon | nodnod | 17:59 |
cjwatson | wgrant: Could you have a look at https://code.launchpad.net/~cjwatson/lazr.restful/py3-declarations/+merge/378548, please? (I think it's probably a bit much to ask people not very familiar with lazr.restful to review that.) | 21:59 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!