CI (rootfs_utils): Require that the user explicitly specify the tag_name

[DilumAluthge]

This will help prevent users from accidentally overwriting an existing tag.

[DilumAluthge]

@staticfloat @tkf Any idea why the asan job is suddenly failing?

[staticfloat]

Nope. Is that what an ASAN error looks like?

[tkf]

I see this first time

julia-debug: /cache/build/amdci7-3/julialang/julia-master/src/gc.c:1201: jl_value_t *jl_gc_pool_alloc(jl_ptls_t, int, int): Assertion `ptls->gc_state == 0' failed.

https://buildkite.com/julialang/julia-master/builds/1911#81137045-be72-4c5a-99db-3c348fa2f20e/1080-1413

The check script also says

┌ Error: Timeout waiting for the subprocess
└ @ Main /cache/build/amdci7-3/julialang/julia-master/contrib/asan/check.jl:76

So I'm not sure if the assertion is due to the kill(proc) from asan/check.jl (though that assertion sounds like it should not get triggered in any case).

I'm building it locally to see if I can reproduce this.

[tkf]

The subprocess spawned from asan/check.jl frist prints the message [ Info: Testing a pattern that would trigger ASAN (example: https://buildkite.com/julialang/julia-master/builds/1905#47de9866-3379-4fa4-8000-03b05ede1eed/1091-1092) but there's no indication of it in this CI log: https://buildkite.com/julialang/julia-master/builds/1911#81137045-be72-4c5a-99db-3c348fa2f20e/1080-1081 Looking at the stacktrace, the error seems to happened around when compiling some ConsoleLogger code.

[tkf]

I can't reproduce this locally.

[DilumAluthge]

Could it be that the subprocess is getting killed by the OOM killer?

[staticfloat]

Nope, dmesg confirms it's a legit segfault:

2021-07-13T19:59:52,330727-0400 julia[29321]: segfault at 1bb8 ip 00007f50a57ce1ae sp 00007ffe77fd6b30 error 4 in libjulia-internal.so.1.8[7f50a56f6000+1cc000]
2021-07-13T19:59:52,362264-0400 julia[29345]: segfault at 1bb8 ip 00007f658f7961ae sp 00007fff0e48e0f0 error 4 in libjulia-internal.so.1.8[7f658f6be000+1cc000]
2021-07-13T19:59:55,698985-0400 julia[30490]: segfault at 7fc738006000 ip 00007fc6f00612d3 sp 00007fff1fac1df0 error 4

[staticfloat]

It's clearly unrelated to this job, but we should keep an eye on it.