[Pytorch] MacOS job failed to clean up disk space before running

[kurtamohler]

A MacOS 15 workflow failed to clean up disk space before running tests: https://github.com/pytorch/pytorch/actions/runs/17473760565/job/49635306741

I thought maybe this was a fluke so I tried rerunning the job, but that didn't fix it.

I also just found a recent commit on the main branch that failed for the same reason: https://github.com/pytorch/pytorch/actions/runs/17478299822/job/49643766504

[benjaminglass1]

I have another failure of this job at https://github.com/pytorch/pytorch/actions/runs/17559538407/job/49873264252

[huydhn]

Assign to myself as OSS oncall this week to root cause this issue. This might come from the new aot_inductor_perf_smoketest on MPS pytorch/pytorch#162153

[malfet]

Please note that it was mentioned in internal workchat on Friday, but not sure what further actions were taken

[huydhn]

I think the issue has been fixed by pytorch/pytorch#162503 as I haven't seen another case in the past few days. Let me monitor this till the end of the week and will resolve it if there is no more failure