Skip to content

fix: port already in use & nccl errors in gpu tests #21335

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 6 commits into from
Nov 6, 2025
+12 −1,208
Merged

fix: port already in use & nccl errors in gpu tests #21335

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
1fb08b5
update
deependujha Nov 5, 2025
f4608dc
update with correct link
bhimrazy Nov 5, 2025
156b8ce
Apply suggestion from @deependujha
deependujha Nov 5, 2025
4efe570
Revert "Fix EADDRINUSE errors in distributed tests with port manager …
bhimrazy Nov 6, 2025
9ac7f92
Merge branch 'revert-21309-fix/port-manager-eaddrinuse' into fix/port…
deependujha Nov 6, 2025
4927768
claymore
deependujha Nov 6, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion .lightning/workflows/fabric.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -140,7 +140,7 @@ run: |

echo "Testing: fabric standalone"
export PL_RUN_STANDALONE_TESTS=1
python -m wget https://raw.githubusercontent.com/Lightning-AI/utilities/main/scripts/run_standalone_tests.sh
python -m wget https://github.com/deependujha/utilities/blob/feat/specify-standalone-port/scripts/run_standalone_tests.sh
bash ./run_standalone_tests.sh "tests_fabric"
export PL_RUN_STANDALONE_TESTS=0

Expand Down
2 changes: 1 addition & 1 deletion .lightning/workflows/pytorch.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -158,7 +158,7 @@ run: |
echo "Testing: fabric standalone"
export PL_USE_MOCKED_MNIST=1
export PL_RUN_STANDALONE_TESTS=1
python -m wget https://raw.githubusercontent.com/Lightning-AI/utilities/main/scripts/run_standalone_tests.sh
python -m wget https://raw.githubusercontent.com/deependujha/utilities/refs/heads/feat/specify-standalone-port/scripts/run_standalone_tests.sh
bash ./run_standalone_tests.sh "tests_pytorch"
export PL_RUN_STANDALONE_TESTS=0

Expand Down
8 changes: 6 additions & 2 deletions src/lightning/fabric/plugins/environments/lightning.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -126,10 +126,14 @@ def find_free_network_port() -> int:
"""
# If an external launcher already specified a MASTER_PORT (for example, torch.distributed.spawn or
# multiprocessing helpers), reserve it through the port manager so no other test reuses the same number.
_port = None
if "STANDALONE_PORT" in os.environ:
_port = os.environ["STANDALONE_PORT"]
if "MASTER_PORT" in os.environ:
master_port_str = os.environ["MASTER_PORT"]
_port = os.environ["MASTER_PORT"]
if _port is not None:
try:
existing_port = int(master_port_str)
existing_port = int(_port)
except ValueError:
pass
else:
Expand Down
Loading