ci: increase free disk space on macos-26 jobs by ubiratansoares · Pull Request #165148 · rust-lang/rust

ubiratansoares · 2026-10-29T17:20:36Z

AI/LLM disclosure

I wrote changes introduced by this single commit by hand. I used an LLM to ideate about changes, and also to review the code before submitting. I used both gpt6/astra and claude 5.5/opus as assistants.

What

A potential fix for rust-lang/infra-team#307 and a follow-up for #165123.

Here we introduce a dedicated script to free disk space on macOS Github runners. We do minimal changes to invoke this script accordingly from the CI Workflow.

The new script works by deleting iOS simulator instance files and runtimes. I assumed these components are not needed and therefore, safe to delete.

I ran a few related experiments in this repo. For macos-26 this script reclaims ~ 120GB of space. That repo also tests most of free-disk-space-mac.sh behavior.

I added a waiting logic after xcrun simctl runtime delete all (which runs asyncly) because in general this command finishes fast (less than 30 seconds) and the script can correctly calculate how much disk space was freed afterwards.

I did not try to replicate the same funcionality available in the existing Linux couterpart of this script, e.g. earlier finishing if some amount of space is already available. I assume that can be a follow-up.

r? marcoieni

ubiratansoares · 2026-10-29T17:34:20Z

@bors try jobs=test-aarch64-apple-macos-26-*

ci: increase free disk space on macos-26 jobs try-job: test-aarch64-apple-macos-26-*

madsmtm

I think the impl is right, though it'll likely have to be updated once we switch to macOS 27 (since there it's xcrun devicectl IIRC).

(I'm not on t-infra though, so you should get one of them to look over it too).

View changes since this review

madsmtm · 2026-10-29T18:06:22Z

+done
+
+if [[ "$remaining" != 0 ]]; then
+    echo "::warning::Runtime deletion did not finish within the polling period."


Nit:

Suggested change

echo "::warning::Runtime deletion did not finish within the polling period."

echo "::warning::Simlator runtime deletion did not finish within the polling period."

madsmtm · 2026-10-29T18:07:18Z

+script_dir=$(cd -- "$(dirname -- "${BASH_SOURCE[0]}")" && pwd)
+
+# dispatch according to the Github Runner
+case "${RUNNER_OS:-}" in


Nit: Would it make sense to do uname here instead? That isn't GH-specific, so easier to understand IMO.

I think we could use uname, but matching against one of the default GHA env vars does not look terrible.

we also have

rust/src/ci/shared.sh

Line 51 in 92b5b8c

function isLinux {

madsmtm · 2026-10-29T18:10:51Z

+# Wait a max of 60 seconds to ensure the clean-up is done before any subsequential build steps
+for ((attempt = 0; attempt < 30; attempt++)); do
+    remaining=$(xcrun simctl runtime list -j | jq 'length')
+    echo "Remaining runtime iOS simulator images : $remaining"


Nit: It's not just iOS simulator images.

Suggested change

echo "Remaining runtime iOS simulator images : $remaining"

echo "Remaining simulator runtime images : $remaining"

Also goes for the top-level comment.

madsmtm · 2026-10-29T18:13:30Z

+#!/bin/bash
+set -euo pipefail
+
+script_dir=$(cd -- "$(dirname -- "${BASH_SOURCE[0]}")" && pwd)


Nit: My bash-fu isn't strong... But I see that musl.sh uses:

SCRIPT_DIR=$(cd -- "$(dirname -- "$0")" && pwd)

What's the benefit to either? If you have no preference, then I'd prefer $0 just for consistency with our other shell scripts.

I think "${BASH_SOURCE[0]}" works better if for some reason one wants to source this file within a bash environment, but maybe I'm missing something. I did some research and, for the current usage, "$0" and "${BASH_SOURCE[0]}" should be same, hence opted for consistency here

madsmtm · 2026-10-29T18:16:14Z

+xcrun simctl delete all
+
+# This subcommand runs asynchronously
+xcrun simctl runtime delete all


Remark: I'm a little bit hesitant to do this, since I have vague plans for running CI on the simulators... But I guess it's more important that CI actually runs, and I can probably tweak this script in the future to keep a few runtime images around.

Thanks for sharing your concern! Happy to improve this later!

we could also delete other stuff in the future instead of what you need 👍

rust-bors · 2026-10-29T20:54:18Z

☀️ Try build successful (CI)
Build commit: e4c3f34 (e4c3f3434fe554c9689b438242a6bf61bc77618c)
Base parent: 96a18bf (96a18bf5ffa2b24f9cf3d1c731a9eb121c1acad9)

ubiratansoares · 2026-10-29T23:29:12Z

@bors try jobs=test-aarch64-apple-macos-26-*

ci: increase free disk space on macos-26 jobs try-job: test-aarch64-apple-macos-26-*

ubiratansoares · 2026-10-29T23:32:52Z

Note: 2a6ff6c in practice reverts #165123 for the sake of testing with macos-26.

rust-bors · 2026-10-30T02:02:52Z

☀️ Try build successful (CI)
Build commit: eb91e94 (eb91e9401c1dcacec25150f2071a662f3e98d799)
Base parent: 92b5b8c (92b5b8c87c61b20adc9001b41baf587634a74734)

marcoieni · 2026-10-30T05:20:05Z

+xcrun simctl delete all
+
+# This subcommand runs asynchronously
+xcrun simctl runtime delete all


we could also delete other stuff in the future instead of what you need 👍

marcoieni · 2026-10-30T05:28:05Z

Overall looks good, feel free to r=me when you are happy with this 👍

rustbot · 2026-10-30T08:16:39Z

Warning

If you are changing how CI LLVM is built or linked, make sure to bump
src/bootstrap/download-ci-llvm-stamp.

cc @jieyouxu

marcoieni · 2026-10-30T15:10:38Z

Feel free to r=me.
Consider if you want to squash your commits. 👍

ubiratansoares · 2026-10-30T15:34:46Z

Lets move with this :)

@bors r=marcoieni

rust-bors · 2026-10-30T15:34:49Z

📌 Commit d91e6e5 has been tentatively approved by marcoieni

It will be put into the queue for this repository once PR CI succeeds.

…pace, r=marcoieni ci: increase free disk space on macos-26 jobs ## AI/LLM disclosure I wrote changes introduced by this single commit by hand. I used an LLM to ideate about changes, and also to review the code before submitting. I used both gpt6/astra and claude 5.5/opus as assistants. ## What A potential fix for rust-lang/infra-team#307 and a follow-up for rust-lang#165123. Here we introduce a dedicated script to free disk space on macOS Github runners. We do minimal changes to invoke this script accordingly from the CI Workflow. The new script works by deleting iOS simulator instance files and runtimes. I assumed these components are not needed and therefore, safe to delete. I ran a few related experiments [in this repo](https://github.com/ubiratansoares/macos-runners-disk-space-tests/actions/runs/40911961033). For `macos-26` this script reclaims ~ 120GB of space. That repo also tests most of `free-disk-space-mac.sh` behavior. I added a waiting logic after `xcrun simctl runtime delete all` (which runs asyncly) because in general this command finishes fast (less than 30 seconds) and the script can correctly calculate how much disk space was freed afterwards. I did not try to replicate the same funcionality available in the existing Linux couterpart of this script, e.g. earlier finishing if some amount of space is already available. I assume that can be a follow-up. r? marcoieni

…uwer Rollup of 22 pull requests Successful merges: - #164851 (Add behavior tests for `Arc` and `Rc`) - #164989 (rustdoc: Fix performance regression for private documented items) - #165079 (Remove GenericArgRef) - #161781 (Don't emit InlineAsm terminator if input operand is `!`.) - #162754 (optimize large elements for slice rotating) - #163252 (Remove windows-targets dependency from the root workspace) - #163839 (Fix the floating point clobbers of RISC-V `clobber_abi`) - #165104 (Use getrandom instead of /scheme/rand on Redox OS) - #165132 (Implement `advance_back_by` on `Fuse`) - #165142 (Remove the unused generic Graphviz writer) - #165148 (ci: increase free disk space on macos-26 jobs) - #163371 (Don't suggest replacing foreign format specifiers in `concat!` output) - #163802 (fix RegionEq direction in canonical query responses) - #163986 (omit unconditionally invoking crate level limits query) - #164920 (Point at prior token on `let` when expression is expected) - #164987 (use adt_async_drop_tys for nested ADTs in needs_async_drop) - #165061 (Use a fixed-length encoding for hash disambiguators in v0 symbols) - #165114 (use native interpreter impl for funnel shifts intrinsics) - #165119 (remove unnecessary fast path from borrowck) - #165122 (fix MaybeDangling validity logic for ZST at usize::MAX) - #165128 (Move `doc(keyword)` and `doc(attribute)` attributes checks from `rustc_passes` to `rustc_attr_parsing`) - #165129 (Fix ICE in prelude edition lint) Failed merges: - #164981 (Reflection add TypeId::name) > [!WARNING] > This rollup conflicts with pending auto build #165205 and may need to be recreated if the pending build succeeds.

JonathanBrouwer · 2026-10-30T21:05:55Z

@bors p=1

…uwer Rollup of 22 pull requests Successful merges: - #164851 (Add behavior tests for `Arc` and `Rc`) - #164989 (rustdoc: Fix performance regression for private documented items) - #165079 (Remove GenericArgRef) - #161781 (Don't emit InlineAsm terminator if input operand is `!`.) - #162754 (optimize large elements for slice rotating) - #163252 (Remove windows-targets dependency from the root workspace) - #163839 (Fix the floating point clobbers of RISC-V `clobber_abi`) - #165104 (Use getrandom instead of /scheme/rand on Redox OS) - #165132 (Implement `advance_back_by` on `Fuse`) - #165142 (Remove the unused generic Graphviz writer) - #165148 (ci: increase free disk space on macos-26 jobs) - #163371 (Don't suggest replacing foreign format specifiers in `concat!` output) - #163802 (fix RegionEq direction in canonical query responses) - #163986 (omit unconditionally invoking crate level limits query) - #164920 (Point at prior token on `let` when expression is expected) - #164987 (use adt_async_drop_tys for nested ADTs in needs_async_drop) - #165061 (Use a fixed-length encoding for hash disambiguators in v0 symbols) - #165114 (use native interpreter impl for funnel shifts intrinsics) - #165119 (remove unnecessary fast path from borrowck) - #165122 (fix MaybeDangling validity logic for ZST at usize::MAX) - #165128 (Move `doc(keyword)` and `doc(attribute)` attributes checks from `rustc_passes` to `rustc_attr_parsing`) - #165129 (Fix ICE in prelude edition lint) Failed merges: - #164981 (Reflection add TypeId::name) > [!WARNING] > This rollup conflicts with pending auto build #165205 and may need to be recreated if the pending build succeeds.

JonathanBrouwer · 2026-10-31T13:32:27Z

@bors r-
Does not seem to fix the issue, CI still doesn't work :c
We should still merge this tho, after we undo #165233

rust-bors · 2026-10-31T13:32:35Z

This pull request was unapproved.

This PR was contained in a rollup (#165208), which was unapproved.

View changes since this unapproval

jieyouxu · 2026-11-01T09:27:23Z

+      RUST_CONFIGURE_ARGS: >-
+        --enable-sanitizers
+        --enable-profiler
+        --set rust.jemalloc


This should use --set build.allocator=jemalloc, ditto for other instances (rust.jemalloc is the old deprecated version)

View changes since the review

jieyouxu · 2026-11-01T09:29:22Z

Non-blocking: should the mac equivalent also have an early exit if there's already estimated sufficient disk space?

rust/src/ci/scripts/free-disk-space-linux.sh

Lines 318 to 326 in eec3f1e

sufficientSpaceEarlyExit() {

local available_space_kb

available_space_kb=$(df -k . --output=avail | tail -n 1)

if [ "$available_space_kb" -ge "$space_target_kb" ]; then

echo "Sufficient disk space available (${available_space_kb}KB >= ${space_target_kb}KB). Skipping cleanup."

exit 0

fi

}

View changes since the review

JonathanBrouwer · 2026-11-01T09:32:37Z

These commits were cherry-picked into #165251

… r=jieyouxu Restore aarch64-apple-1 jobs with free disk space The `apple-1` jobs seem to work with the free disk space script that I've copied from rust-lang#165148 Lets re-enable them The commits are cherry-picked from rust-lang#165148 without changes, plus a few added commits to disable the `-2` jobs again. r? ghost

Rollup merge of #165251 - JonathanBrouwer:undo-apple-remove, r=jieyouxu Restore aarch64-apple-1 jobs with free disk space The `apple-1` jobs seem to work with the free disk space script that I've copied from #165148 Lets re-enable them The commits are cherry-picked from #165148 without changes, plus a few added commits to disable the `-2` jobs again. r? ghost

rustbot assigned marcoieni Oct 29, 2026

rust-bors Bot pushed a commit that referenced this pull request Oct 29, 2026

Auto merge of #165148 - ubiratansoares:u/macos-free-disk-space, r=<try>

e4c3f34

ci: increase free disk space on macos-26 jobs try-job: test-aarch64-apple-macos-26-*

This comment has been minimized.

Sign in to view

madsmtm approved these changes Oct 29, 2026

View reviewed changes

This comment has been minimized.

Sign in to view

ubiratansoares added 3 commits October 30, 2026 00:23

ci: increase free disk space on macos-26 jobs

fb5a8e9

ci: add minor improvements to scripts

aa92deb

ci: restore macos-26 jobs

2a6ff6c

ubiratansoares force-pushed the u/macos-free-disk-space branch from f35f96e to 2a6ff6c Compare October 29, 2026 23:27

This comment has been minimized.

Sign in to view

rust-bors Bot pushed a commit that referenced this pull request Oct 29, 2026

Auto merge of #165148 - ubiratansoares:u/macos-free-disk-space, r=<try>

eb91e94

ci: increase free disk space on macos-26 jobs try-job: test-aarch64-apple-macos-26-*

marcoieni reviewed Oct 30, 2026

View reviewed changes

ci: output df stats before and after clean-up

356f23f

ubiratansoares marked this pull request as ready for review October 30, 2026 08:16

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Oct 30, 2026

ci: free disk space also on macos-15

d91e6e5

rust-bors Bot added S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Oct 30, 2026

JonathanBrouwer mentioned this pull request Oct 30, 2026

Rollup of 22 pull requests #165208

Closed

rust-bors Bot added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. labels Oct 31, 2026

jieyouxu reviewed Nov 1, 2026

View reviewed changes

JonathanBrouwer closed this Nov 1, 2026

rustbot removed the S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. label Nov 1, 2026

JonathanBrouwer mentioned this pull request Nov 1, 2026

Restore aarch64-apple-1 jobs with free disk space #165251

Merged

	echo "::warning::Runtime deletion did not finish within the polling period."
	echo "::warning::Simlator runtime deletion did not finish within the polling period."

	echo "Remaining runtime iOS simulator images : $remaining"
	echo "Remaining simulator runtime images : $remaining"

	sufficientSpaceEarlyExit() {
	local available_space_kb
	available_space_kb=$(df -k . --output=avail \| tail -n 1)

	if [ "$available_space_kb" -ge "$space_target_kb" ]; then
	echo "Sufficient disk space available (${available_space_kb}KB >= ${space_target_kb}KB). Skipping cleanup."
	exit 0
	fi
	}

Conversation

ubiratansoares commented Oct 29, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

AI/LLM disclosure

What

Uh oh!

ubiratansoares commented Oct 29, 2026

Uh oh!

This comment has been minimized.

madsmtm left a comment • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ubiratansoares Oct 29, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ubiratansoares Oct 29, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ubiratansoares Oct 29, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

rust-bors Bot commented Oct 29, 2026

Uh oh!

This comment has been minimized.

ubiratansoares commented Oct 29, 2026

Uh oh!

This comment has been minimized.

ubiratansoares commented Oct 29, 2026

Uh oh!

rust-bors Bot commented Oct 30, 2026

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

marcoieni commented Oct 30, 2026

Uh oh!

rustbot commented Oct 30, 2026

Uh oh!

marcoieni commented Oct 30, 2026

Uh oh!

ubiratansoares commented Oct 30, 2026

Uh oh!

rust-bors Bot commented Oct 30, 2026

Uh oh!

JonathanBrouwer commented Oct 30, 2026

Uh oh!

JonathanBrouwer commented Oct 31, 2026

Uh oh!

rust-bors Bot commented Oct 31, 2026

Uh oh!

jieyouxu Nov 1, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

ubiratansoares commented Oct 29, 2026 •

edited by rustbot

Loading

madsmtm left a comment •

edited by rustbot

Loading

ubiratansoares Oct 29, 2026 •

edited

Loading

ubiratansoares Oct 29, 2026 •

edited

Loading

ubiratansoares Oct 29, 2026 •

edited

Loading

jieyouxu Nov 1, 2026 •

edited by rustbot

Loading

jieyouxu Nov 1, 2026 •

edited by rustbot

Loading