[zstd] CIFuzz failing to build

[terrelln]

The script build.sh started failing a few days ago on this line.

oss-fuzz/projects/zstd/build.sh

Line 19 in 600c514

cd tests/fuzz

It looks like build.sh assumes it is being run in the root of the zstd repo. But, some time ago it stopped being run from the root of the zstd repo. Is that expected?

[terrelln]

E.g. this run https://github.com/facebook/zstd/pull/2231/checks?check_run_id=849291310

[Dor1s]

As I recall CIFuzz might've had some assumptions. @jonathanmetzman could you please take a look?

[rgacogne]

CIFuzz stopped working for PowerDNS as well. We could change our working directory in build.sh to ${SRC}/pdns/ but I assume quite a lot of projects will have the same issue so perhaps it would be better to ensure that build.sh is run from the same place it used to, and match the regular OSS-Fuzz behaviour?

[inferno-chromium]

Looks like all cifuzz projects are broken, mayday!
Failures seem around 2020-07-08 08:46:00 UTC

[inferno-chromium]

Failure run:
http://cifuzz.appspot.com/run/pdns/161705947
2020-07-08T08:49:48.4007782Z Running: docker run --rm --privileged --cap-add SYS_PTRACE -e FUZZING_ENGINE=libfuzzer -e SANITIZER=address -e ARCHITECTURE=x86_64 -e FUZZING_LANGUAGE=c++ -e OUT=/out -v /github/workspace/storage/pdns:/src/pdns -v /github/workspace/out:/out gcr.io/oss-fuzz/powerdns /bin/bash -c compile

Success run:
http://cifuzz.appspot.com/run/pdns/160774634
docker run --rm --privileged --cap-add SYS_PTRACE -e FUZZING_ENGINE=libfuzzer -e SANITIZER=address -e ARCHITECTURE=x86_64 -e FUZZING_LANGUAGE=c++ -e OUT=/github/workspace/out --volumes-from 16f1c38a2c11 gcr.io/oss-fuzz/powerdns /bin/bash -c 'rm -rf /src/pdns/* && cp -r /github/workspace/storage/pdns /src && compile'

Looks like utils.get_container_name() helper is broken, looking more.

Maybe something broke with how we are detecting containers
https://github.com/google/oss-fuzz/blob/master/infra/cifuzz/cifuzz.py#L156

[evverx]

Looks like all cifuzz projects are broken, mayday!

FWIW systemd and util-linux are fine

Maybe something broke with how we are detecting containers

It appears some GitHub actions are run with docker.service (where docker is in /proc/self/cgroup) while the zstd actions are run with containerd.service where /proc/self/cgroup looks like

12:rdma:/
11:hugetlb:/actions_job/406ac3d3609a64ae01b084272f02ca57f2a807d1db3f3bd32c0c3f478f515262
10:cpu,cpuacct:/actions_job/406ac3d3609a64ae01b084272f02ca57f2a807d1db3f3bd32c0c3f478f515262
9:freezer:/actions_job/406ac3d3609a64ae01b084272f02ca57f2a807d1db3f3bd32c0c3f478f515262
8:memory:/actions_job/406ac3d3609a64ae01b084272f02ca57f2a807d1db3f3bd32c0c3f478f515262
7:net_cls,net_prio:/actions_job/406ac3d3609a64ae01b084272f02ca57f2a807d1db3f3bd32c0c3f478f515262
6:perf_event:/actions_job/406ac3d3609a64ae01b084272f02ca57f2a807d1db3f3bd32c0c3f478f515262
5:cpuset:/actions_job/406ac3d3609a64ae01b084272f02ca57f2a807d1db3f3bd32c0c3f478f515262
4:devices:/actions_job/406ac3d3609a64ae01b084272f02ca57f2a807d1db3f3bd32c0c3f478f515262
3:blkio:/actions_job/406ac3d3609a64ae01b084272f02ca57f2a807d1db3f3bd32c0c3f478f515262
2:pids:/actions_job/406ac3d3609a64ae01b084272f02ca57f2a807d1db3f3bd32c0c3f478f515262
1:name=systemd:/actions_job/406ac3d3609a64ae01b084272f02ca57f2a807d1db3f3bd32c0c3f478f515262
0::/system.slice/containerd.service

I've prepared a patch where systemd-detect-virt -c is called to make sure CIFuzz is run in a docker container. It's being tested in facebook/zstd#2239

[inferno-chromium]

Thanks a ton @evverx for your awesome debugging!