ceos/unsloth
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Harden flaky Studio CI: retry VS-hide rename and tolerate same-URL nav interrupt (#6713)

Browse source 
Two intermittent Studio CI failures, both runner-environment flakes unrelated
to test logic:

Windows 'Studio install + inference without Visual Studio': the 'Hide Visual
Studio + CMake' step renames C:\Program Files\Microsoft Visual Studio to
simulate a host with no build tools. A background handle on a Program Files
directory (Defender scan or an MSBuild node) makes Rename-Item intermittently
fail with 'Access is denied', and $ErrorActionPreference = Stop turns that into
a hard job failure. Wrap the VS and cmake renames in both Hide steps in a short
Rename-WithRetry (6 tries, 3s apart) to ride out the transient lock.

macOS 'Chat UI Tests': the re-login goto to /login can be interrupted by the
SPA auth guard redirecting to the same /login URL, which Playwright reports as
'Navigation to .../login is interrupted by another navigation to .../login'.
The goto already tolerated ERR_ABORTED; broaden it to also tolerate the same-URL
interrupt (the password-field wait right after confirms we landed on /login),
and add the same signature to the two Playwright flake-retry harnesses as a
safety net for any other navigation.

Validated: playwright_chat_ui.py parses + byte-compiles, both workflow YAMLs
parse, bash -n on the retry harnesses, PowerShell AST parse on all pwsh steps,
and a functional check of Rename-WithRetry (succeeds, and rethrows after
exhausting retries).
This commit is contained in:
Daniel Han 2026-06-26 19:45:46 -07:00 • committed by GitHub
commit 1fcd69e662
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
3 changed files with 42 additions and 21 deletions
Download patch file Download diff file Expand all files Collapse all files

29
.github/workflows/studio-mac-ui-smoke.yml vendored
View file

@ -185,13 +185,14 @@ jobs:
# Retry up to 3 times to absorb known macos-14 free-runner
# flakes: (1) Playwright Node 24 pipeTransport.js 'Unexpected
# end of JSON input' crash when the Chromium browser process
# dies mid-test, and (2) Chromium net::ERR_NO_BUFFER_SPACE
# when the runner's kernel briefly runs out of socket buffers.
# The retry FULLY resets Studio (kill, reset-password, reboot,
# wait /api/health, re-export bootstrap pw) before re-running
# the script. A real test failure (assertion / timeout) does
# NOT match either pattern so it bypasses retry and surfaces
# immediately.
# dies mid-test, (2) Chromium net::ERR_NO_BUFFER_SPACE when the
# runner's kernel briefly runs out of socket buffers, and (3) a
# goto 'interrupted by another navigation' when the SPA auth
# guard redirects mid-navigation. The retry FULLY resets Studio
# (kill, reset-password, reboot, wait /api/health, re-export
# bootstrap pw) before re-running the script. A real test failure
# (assertion / timeout) does NOT match any pattern so it bypasses
# retry and surfaces immediately.
run: |
mkdir -p logs/playwright
attempt=1
@ -204,8 +205,9 @@ jobs:
if [ "$rc" -eq 0 ]; then
break
fi
if { grep -q "Unexpected end of JSON input" logs/playwright_attempt_${attempt}.log \
|| grep -q "ERR_NO_BUFFER_SPACE" logs/playwright_attempt_${attempt}.log; } \
if { grep -q "Unexpected end of JSON input" logs/playwright_attempt_${attempt}.log \
|| grep -q "ERR_NO_BUFFER_SPACE" logs/playwright_attempt_${attempt}.log \
|| grep -q "interrupted by another navigation" logs/playwright_attempt_${attempt}.log; } \
&& [ "$attempt" -lt "$max_attempts" ]; then
echo "::warning::Playwright flake on attempt ${attempt}; resetting Studio and retrying..."
kill "${STUDIO_PID}" 2>/dev/null || true
@ -280,8 +282,8 @@ jobs:
STUDIO_UI_TURN_TIMEOUT_MS: '540000'
GGUF_REPO: ${{ env.GGUF_REPO }}
GGUF_VARIANT: ${{ env.GGUF_VARIANT }}
# Same flake-retry shape as "Drive the chat UI with Playwright"
# -- catches pipeTransport JSON crash and ERR_NO_BUFFER_SPACE.
# Same flake-retry shape as "Drive the chat UI with Playwright" -- catches
# pipeTransport JSON crash, ERR_NO_BUFFER_SPACE, and nav interrupts.
run: |
mkdir -p logs/playwright_extra
attempt=1
@ -294,8 +296,9 @@ jobs:
if [ "$rc" -eq 0 ]; then
break
fi
if { grep -q "Unexpected end of JSON input" logs/playwright_extra_attempt_${attempt}.log \
|| grep -q "ERR_NO_BUFFER_SPACE" logs/playwright_extra_attempt_${attempt}.log; } \
if { grep -q "Unexpected end of JSON input" logs/playwright_extra_attempt_${attempt}.log \
|| grep -q "ERR_NO_BUFFER_SPACE" logs/playwright_extra_attempt_${attempt}.log \
|| grep -q "interrupted by another navigation" logs/playwright_extra_attempt_${attempt}.log; } \
&& [ "$attempt" -lt "$max_attempts" ]; then
echo "::warning::Playwright flake on attempt ${attempt}; resetting Studio and retrying..."
kill "${STUDIO_EXTRA_PID}" 2>/dev/null || true

22
.github/workflows/studio-windows-inference-smoke.yml vendored
View file

@ -1338,11 +1338,19 @@ jobs:
shell: pwsh
run: |
$ErrorActionPreference = 'Stop'
# A Program Files dir can hold a transient handle (Defender / MSBuild node)
# so Rename-Item intermittently fails with "Access is denied"; retry to ride it out.
function Rename-WithRetry($Path, $NewName) {
for ($i = 1; $i -le 6; $i++) {
try { Rename-Item -LiteralPath $Path -NewName $NewName -ErrorAction Stop; return }
catch { if ($i -eq 6) { throw }; Start-Sleep -Seconds 3 }
}
}
# Rename the Visual Studio install roots (incl. the Installer that holds
# vswhere.exe) so Find-VsBuildTools' vswhere + filesystem scan both miss.
foreach ($d in @("$env:ProgramFiles\Microsoft Visual Studio", "${env:ProgramFiles(x86)}\Microsoft Visual Studio")) {
if (Test-Path -LiteralPath $d) {
Rename-Item -LiteralPath $d -NewName ((Split-Path $d -Leaf) + '.vsoff')
Rename-WithRetry $d ((Split-Path $d -Leaf) + '.vsoff')
Write-Host "Hid VS: $d"
}
}
@ -1351,7 +1359,7 @@ jobs:
$hidden = @()
foreach ($c in (Get-Command cmake -All -ErrorAction SilentlyContinue)) {
if ($c.Source -and (Test-Path -LiteralPath $c.Source)) {
Rename-Item -LiteralPath $c.Source -NewName ((Split-Path $c.Source -Leaf) + '.off')
Rename-WithRetry $c.Source ((Split-Path $c.Source -Leaf) + '.off')
$hidden += $c.Source
Write-Host "Hid cmake: $($c.Source)"
}
@ -1536,8 +1544,16 @@ jobs:
shell: pwsh
run: |
$ErrorActionPreference = 'Stop'
# Retry the rename: a Program Files dir can hold a transient handle that
# makes Rename-Item intermittently fail with "Access is denied".
function Rename-WithRetry($Path, $NewName) {
for ($i = 1; $i -le 6; $i++) {
try { Rename-Item -LiteralPath $Path -NewName $NewName -ErrorAction Stop; return }
catch { if ($i -eq 6) { throw }; Start-Sleep -Seconds 3 }
}
}
foreach ($d in @("$env:ProgramFiles\Microsoft Visual Studio", "${env:ProgramFiles(x86)}\Microsoft Visual Studio")) {
if (Test-Path -LiteralPath $d) { Rename-Item -LiteralPath $d -NewName ((Split-Path $d -Leaf) + '.vsoff'); Write-Host "Hid VS: $d" }
if (Test-Path -LiteralPath $d) { Rename-WithRetry $d ((Split-Path $d -Leaf) + '.vsoff'); Write-Host "Hid VS: $d" }
}
- name: Windows CUDA and ROCm prebuilts exist in unslothai/llama.cpp (what GPU users download, no VS)

12
tests/studio/playwright_chat_ui.py
View file

@ -1225,15 +1225,17 @@ with sync_playwright() as p:
# ─────────────────────────────────────────────────────
step("Shutdown via account menu")
# Re-login with NEW2 for a valid /api/shutdown token (CLI rotation
# invalidated the old one). The stale token can make the SPA auth
# guard abort this goto with ERR_ABORTED; resolve on
# domcontentloaded and tolerate it -- the pw-field wait confirms /login.
# invalidated the old one). The stale token can make the SPA auth guard
# abort this goto with ERR_ABORTED, or redirect to the same /login URL
# ("interrupted by another navigation"); resolve on domcontentloaded and
# tolerate either -- the pw-field wait below confirms we are on /login.
_tolerated_nav = ("ERR_ABORTED", "interrupted by another navigation")
try:
page.goto(f"{BASE}/login", wait_until = "domcontentloaded", timeout = 60_000)
except Exception as exc:
if "ERR_ABORTED" not in str(exc):
if not any(t in str(exc) for t in _tolerated_nav):
raise
info(f"goto /login aborted ({exc!r}); password-field wait will confirm /login")
info(f"goto /login interrupted ({exc!r}); password-field wait will confirm /login")
pw_field = page.locator("#password")
pw_field.wait_for(state = "visible", timeout = 60_000)
pw_field.fill(NEW2)