testutil: pin busybox and alpine used in releases

[tonistiigi]

replaces #3464

Signed-off-by: Tonis Tiigi tonistiigi@gmail.com

[AkihiroSuda]

Still failing https://github.com/moby/buildkit/actions/runs/3858054800/jobs/6576171616

#5 [build 1/3] FROM docker.io/library/busybox@sha256:7b3ccabffc97de872a30dfd234fd972a66d247c8cfc69b0550f276481852627c
#5 resolve docker.io/library/busybox@sha256:7b3ccabffc97de872a30dfd234fd972a66d247c8cfc69b0550f276481852627c done
#5 sha256:205dae5015e78dd8c4d302e3db4eb31576fac715b46d099fe09680ba28093a7a 2.59MB / 2.59MB 0.1s done
#5 extracting sha256:205dae5015e78dd8c4d302e3db4eb31576fac715b46d099fe09680ba28093a7a 0.1s done
#5 DONE 0.2s

#6 [build 2/3] RUN cat /dev/urandom | head -c 100 | sha256sum > unique_first
#6 0.479 Illegal instruction (core dumped)
time="2023-01-06T19:56:23Z" level=error msg="/moby.buildkit.v1.Control/Solve returned error: rpc error: code = Unknown desc = process \"/bin/sh -c cat /dev/urandom | head -c 100 | sha256sum > unique_first\" did not complete successfully: exit code: 132"
#6 ERROR: process "/bin/sh -c cat /dev/urandom | head -c 100 | sha256sum > unique_first" did not complete successfully: exit code: 132
------
 > [build 2/3] RUN cat /dev/urandom | head -c 100 | sha256sum > unique_first:
#6 0.479 Illegal instruction (core dumped)
------
Dockerfile:2
--------------------
   1 |     FROM busybox AS build
   2 | >>> RUN cat /dev/urandom | head -c 100 | sha256sum > unique_first
   3 |     RUN cat /dev/urandom | head -c 100 | sha256sum > unique_second
   4 |     
--------------------
error: failed to solve: process "/bin/sh -c cat /dev/urandom | head -c 100 | sha256sum > unique_first" did not complete successfully: exit code: 132

[AkihiroSuda]

Can we just use the digest of the multi-platform index?

[tonistiigi]

iirc this was optimization in amd64/arm64 to avoid mirroring the unneeded architectures. Other arch use library ns with multi-platform.

[AkihiroSuda]

Is this issue tracked somewhere?

Can't find a ticket in https://github.com/docker-library/busybox/issues and https://bugs.busybox.net/buglist.cgi?chfield=%5BBug%20creation%5D&chfieldfrom=30 .

Seems related to mirror/busybox@6472ac9 though.

[tonistiigi]

I haven't reported to busybox yet. If you do, you can post this debug: https://github.com/tonistiigi/gh-busybox-sigill-debug/actions/runs/3858482123

Good CPUs: 8370C, E5-2673
Bad CPUs: 8171M, 8272CL

[tonistiigi]

https://bugs.busybox.net/show_bug.cgi?id=15236