Skip to content

[Log](pyudf) Add progress logs for python process pool init #62974

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

Open
linrrzqqq wants to merge 2 commits into
base: master
Choose a base branch
from
Open

[Log](pyudf) Add progress logs for python process pool init #62974

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
591ac87
[Log](pyudf) Add progress logs for python process pool init
linrrzqqq Apr 30, 2026
076ba81
fix beut
linrrzqqq Apr 30, 2026
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
23 changes: 22 additions & 1 deletion be/src/udf/python/python_server.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,7 +26,10 @@

#include <boost/asio.hpp>
#include <boost/process.hpp>
#include <chrono>
#include <fstream>
#include <future>
#include <thread>

#include "arrow/flight/client.h"
#include "common/config.h"
Expand Down Expand Up @@ -88,7 +91,21 @@ Status PythonServerManager::ensure_pool_initialized(const PythonVersion& version

int success_count = 0;
int failure_count = 0;
const auto init_start_time = std::chrono::steady_clock::now();
constexpr auto progress_log_interval = std::chrono::seconds(20);
for (int i = 0; i < max_pool_size; i++) {
// Print init log every 20s until the current slot is ready.
while (futures[i].wait_for(progress_log_interval) != std::future_status::ready) {
const auto now = std::chrono::steady_clock::now();
const auto total_elapsed_ms =
std::chrono::duration_cast<std::chrono::milliseconds>(now - init_start_time)
.count();
LOG(INFO) << "Python process pool initialization progress for version "
<< version.to_string() << ": waiting_slot=" << (i + 1) << "/" << max_pool_size
<< ", success=" << success_count << ", failed=" << failure_count
<< ", elapsed_ms=" << total_elapsed_ms;
}

Status s = futures[i].get();
if (s.ok() && temp_processes[i]) {
pool.push_back(std::move(temp_processes[i]));
Expand All @@ -106,9 +123,13 @@ Status PythonServerManager::ensure_pool_initialized(const PythonVersion& version
max_pool_size);
}

const auto total_elapsed_ms = std::chrono::duration_cast<std::chrono::milliseconds>(
std::chrono::steady_clock::now() - init_start_time)
.count();
LOG(INFO) << "Python process pool initialized for version " << version.to_string()
<< ": created " << success_count << " processes"
<< (failure_count > 0 ? fmt::format(" ({} failed)", failure_count) : "");
<< (failure_count > 0 ? fmt::format(" ({} failed)", failure_count) : "")
<< ", elapsed_ms=" << total_elapsed_ms;

_initialized_versions.insert(version);
_start_health_check_thread();
Expand Down
29 changes: 15 additions & 14 deletions be/test/udf/python/python_udf_runtime_test.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -205,9 +205,10 @@ TEST_F(PythonUDFRuntimeTest, FlightServerPathTemplate) {
// ============================================================================

TEST_F(PythonUDFRuntimeTest, ShutdownTerminatesProcess) {
// Use /bin/cat which blocks waiting for stdin - reliable and fast
// Use sleep instead of a stdin-driven command. In CI, stdin may be closed and
// commands like cat can exit before running() is checked.
bp::ipstream output;
bp::child child("/bin/cat", bp::std_out > output);
bp::child child("/bin/sleep", "60", bp::std_out > output);

ASSERT_TRUE(child.valid());
ASSERT_TRUE(child.running());
Expand All @@ -229,7 +230,7 @@ TEST_F(PythonUDFRuntimeTest, ShutdownTerminatesProcess) {

TEST_F(PythonUDFRuntimeTest, ShutdownIdempotent) {
bp::ipstream output;
bp::child child("/bin/cat", bp::std_out > output);
bp::child child("/bin/sleep", "60", bp::std_out > output);

PythonUDFProcess process(std::move(child), std::move(output));

Expand All @@ -247,7 +248,7 @@ TEST_F(PythonUDFRuntimeTest, ShutdownIdempotent) {
TEST_F(PythonUDFRuntimeTest, ShutdownWithStubbornProcess) {
// Create a process that ignores SIGTERM - tests the SIGKILL fallback path
bp::ipstream output;
bp::child child("/bin/bash", "-c", "trap '' TERM; cat", bp::std_out > output);
bp::child child("/bin/bash", "-c", "trap '' TERM; exec sleep 60", bp::std_out > output);

PythonUDFProcess process(std::move(child), std::move(output));
EXPECT_TRUE(process.is_alive());
Expand All @@ -265,7 +266,7 @@ TEST_F(PythonUDFRuntimeTest, ShutdownWithStubbornProcess) {

TEST_F(PythonUDFRuntimeTest, RemoveUnixSocketExistingFile) {
bp::ipstream output;
bp::child child("/bin/cat", bp::std_out > output);
bp::child child("/bin/sleep", "60", bp::std_out > output);

pid_t child_pid = child.id();
PythonUDFProcess process(std::move(child), std::move(output));
Expand All @@ -284,7 +285,7 @@ TEST_F(PythonUDFRuntimeTest, RemoveUnixSocketExistingFile) {

TEST_F(PythonUDFRuntimeTest, RemoveUnixSocketNonExistent) {
bp::ipstream output;
bp::child child("/bin/cat", bp::std_out > output);
bp::child child("/bin/sleep", "60", bp::std_out > output);

pid_t child_pid = child.id();
PythonUDFProcess process(std::move(child), std::move(output));
Expand All @@ -300,7 +301,7 @@ TEST_F(PythonUDFRuntimeTest, RemoveUnixSocketNonExistent) {

TEST_F(PythonUDFRuntimeTest, RemoveUnixSocketIsDirectory) {
bp::ipstream output;
bp::child child("/bin/cat", bp::std_out > output);
bp::child child("/bin/sleep", "60", bp::std_out > output);

pid_t child_pid = child.id();

Expand All @@ -326,7 +327,7 @@ TEST_F(PythonUDFRuntimeTest, RemoveUnixSocketIsDirectory) {

TEST_F(PythonUDFRuntimeTest, ToStringFormat) {
bp::ipstream output;
bp::child child("/bin/cat", bp::std_out > output);
bp::child child("/bin/sleep", "60", bp::std_out > output);

pid_t child_pid = child.id();
PythonUDFProcess process(std::move(child), std::move(output));
Expand All @@ -350,7 +351,7 @@ TEST_F(PythonUDFRuntimeTest, ToStringFormat) {

TEST_F(PythonUDFRuntimeTest, GetUri) {
bp::ipstream output;
bp::child child("/bin/cat", bp::std_out > output);
bp::child child("/bin/sleep", "60", bp::std_out > output);

pid_t child_pid = child.id();
PythonUDFProcess process(std::move(child), std::move(output));
Expand All @@ -367,7 +368,7 @@ TEST_F(PythonUDFRuntimeTest, GetUri) {

TEST_F(PythonUDFRuntimeTest, GetSocketFilePath) {
bp::ipstream output;
bp::child child("/bin/cat", bp::std_out > output);
bp::child child("/bin/sleep", "60", bp::std_out > output);

pid_t child_pid = child.id();
PythonUDFProcess process(std::move(child), std::move(output));
Expand All @@ -389,8 +390,8 @@ TEST_F(PythonUDFRuntimeTest, GetSocketFilePath) {

TEST_F(PythonUDFRuntimeTest, ProcessEquality) {
bp::ipstream output1, output2;
bp::child child1("/bin/cat", bp::std_out > output1);
bp::child child2("/bin/cat", bp::std_out > output2);
bp::child child1("/bin/sleep", "60", bp::std_out > output1);
bp::child child2("/bin/sleep", "60", bp::std_out > output2);

PythonUDFProcess process1(std::move(child1), std::move(output1));
PythonUDFProcess process2(std::move(child2), std::move(output2));
Expand All @@ -410,7 +411,7 @@ TEST_F(PythonUDFRuntimeTest, DestructorCallsShutdown) {
pid_t child_pid;
{
bp::ipstream output;
bp::child child("/bin/cat", bp::std_out > output);
bp::child child("/bin/sleep", "60", bp::std_out > output);
child_pid = child.id();

PythonUDFProcess process(std::move(child), std::move(output));
Expand All @@ -431,7 +432,7 @@ TEST_F(PythonUDFRuntimeTest, DestructorCallsShutdown) {

TEST_F(PythonUDFRuntimeTest, IsAliveReflectsState) {
bp::ipstream output;
bp::child child("/bin/cat", bp::std_out > output);
bp::child child("/bin/sleep", "60", bp::std_out > output);

PythonUDFProcess process(std::move(child), std::move(output));

Expand Down
2 changes: 1 addition & 1 deletion regression-test/pipeline/cloud_p0/conf/be_custom.conf
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -63,7 +63,7 @@ fail_when_segment_rows_not_in_rowset_meta=true
enable_python_udf_support=true
python_env_mode=conda
python_conda_root_path=/opt/miniconda3
max_python_process_num=64
max_python_process_num=16

enable_cloud_make_rs_visible_on_be=true
cloud_mow_sync_rowsets_when_load_txn_begin=false
2 changes: 1 addition & 1 deletion regression-test/pipeline/cloud_p1/conf/be_custom.conf
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -45,7 +45,7 @@ enable_segment_rows_check_core=true
enable_python_udf_support=true
python_env_mode=conda
python_conda_root_path=/opt/miniconda3
max_python_process_num=64
max_python_process_num=16

enable_cloud_make_rs_visible_on_be=true
cloud_mow_sync_rowsets_when_load_txn_begin=false
2 changes: 1 addition & 1 deletion regression-test/pipeline/external/conf/be.conf
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -77,4 +77,4 @@ enable_graceful_exit_check=true
enable_python_udf_support=true
python_env_mode=venv
python_venv_interpreter_paths=/usr/bin/python
max_python_process_num=64
max_python_process_num=16
2 changes: 1 addition & 1 deletion regression-test/pipeline/nonConcurrent/conf/be.conf
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -98,4 +98,4 @@ enable_segment_rows_check_core=true
enable_python_udf_support=true
python_env_mode=venv
python_venv_interpreter_paths=/usr/bin/python
max_python_process_num=64
max_python_process_num=16
2 changes: 1 addition & 1 deletion regression-test/pipeline/p0/conf/be.conf
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -101,4 +101,4 @@ fail_when_segment_rows_not_in_rowset_meta=true
enable_python_udf_support=true
python_env_mode=venv
python_venv_interpreter_paths=/usr/bin/python
max_python_process_num=64
max_python_process_num=16
2 changes: 1 addition & 1 deletion regression-test/pipeline/p1/conf/be.conf
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -84,4 +84,4 @@ enable_segment_rows_check_core=true
enable_python_udf_support=true
python_env_mode=venv
python_venv_interpreter_paths=/usr/bin/python
max_python_process_num=64
max_python_process_num=16
2 changes: 1 addition & 1 deletion regression-test/pipeline/vault_p0/conf/be_custom.conf
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -44,4 +44,4 @@ enable_brpc_connection_check=true
enable_python_udf_support=true
python_env_mode=venv
python_venv_interpreter_paths=/usr/bin/python
max_python_process_num=64
max_python_process_num=16
Loading