Re: Infrastructure | Run CI containers with --cap-add SYS_PTRACE (#370)

2020-07-20 Thread Philip Chimento
Philip Chimento commented: I'm so happy this could be done. It's really going to save me a lot of time, and prevent a lot of frustration from contributors. -- Reply to this email directly or view it on GitLab: https://gitlab.gnome.org/Infrastructure/Infrastructure/-/issues/370#note_870994

Re: Infrastructure | Run CI containers with --cap-add SYS_PTRACE (#370)

2020-07-20 Thread Michael Catanzaro
Michael Catanzaro commented: Thanks! -- Reply to this email directly or view it on GitLab: https://gitlab.gnome.org/Infrastructure/Infrastructure/-/issues/370#note_870856 You're receiving this email because of your account on gitlab.gnome.org.

Re: Infrastructure | Run CI containers with --cap-add SYS_PTRACE (#370)

2020-07-20 Thread Bartłomiej Piotrowski
Bartłomiej Piotrowski commented: @mcatanzaro @pwithnall There are 4 shared runners with `asan` tag which have SYS_PTRACE passed to Docker. I have submitted MRs to glib-networking and gjs enabling them; please poke me somewhere if you notice any problems. -- Reply to this email directly or

Re: Infrastructure | Run CI containers with --cap-add SYS_PTRACE (#370)

2020-07-20 Thread Bartłomiej Piotrowski
Issue was closed by Bartłomiej Piotrowski Issue #370: https://gitlab.gnome.org/Infrastructure/Infrastructure/-/issues/370 -- Reply to this email directly or view it on GitLab: https://gitlab.gnome.org/Infrastructure/Infrastructure/-/issues/370 You're receiving this email because of your

Re: Infrastructure | Run CI containers with --cap-add SYS_PTRACE (#370)

2020-07-13 Thread Sri Ramkrishna
Sri Ramkrishna commented: I'm watching this issue as I'm also working on a ci runner. I can understand that turning off CI for external contributors in World when ci runner resources are limited is a good thing. But I'm hoping that we can still use it for extensions rebooted. If I have to

Re: Infrastructure | Run CI containers with --cap-add SYS_PTRACE (#370)

2020-07-13 Thread Michael Catanzaro
Michael Catanzaro commented: > I find it hard to believe you don't know security implications of granting > `SYS_PTRACE` to CI which runs completely arbitrary loads, not to mention > runners no longer run with `--privileged` after it was reported it breaks > glib test suite as well.

Re: Infrastructure | Run CI containers with --cap-add SYS_PTRACE (#370)

2020-07-13 Thread Bartłomiej Piotrowski
Bartłomiej Piotrowski commented: I find it hard to believe you don't know security implications of granting `SYS_PTRACE` to CI which runs completely arbitrary loads, not to mention runners no longer run with `--privileged` after it was reported it breaks glib test suite as well. We likely

Re: Infrastructure | Run CI containers with --cap-add SYS_PTRACE (#370)

2020-07-06 Thread Michael Catanzaro
Michael Catanzaro commented: BTW, the error message: ``` ==277==LeakSanitizer has encountered a fatal error. ==277==HINT: For debugging, try setting environment variable LSAN_OPTIONS=verbosity=1:log_threads=1 ==277==HINT: LeakSanitizer does not work under ptrace (strace, gdb, etc) ``` --

Infrastructure | Run CI containers with --cap-add SYS_PTRACE (#370)

2020-07-06 Thread Michael Catanzaro
Michael Catanzaro created an issue: https://gitlab.gnome.org/Infrastructure/Infrastructure/-/issues/370 For the past couple of months, we've needed privileged runners to run tests with asan. It seems to be a regression caused by changes to the container system the CI runs in. It means that