ragflow/docker/entrypoint.sh
Zhichang Yu afe9269534
Introduced jemalloc (#5590)
### What problem does this PR solve?

Introduced jemalloc.
Python uses pymalloc (which is an reimplementation of gblibc malloc) to
manage RES. It has pools for small objects to avoid returning memory to
OS aggressively. My experience is: Replacing pymalloc with
[jemalloc](https://github.com/jemalloc/jemalloc) can reduce RES and
speedup task_executor.py.

### Type of change

- [x] Performance Improvement
2025-03-04 12:49:39 +08:00

36 lines
759 B
Bash
Executable File

#!/bin/bash
# replace env variables in the service_conf.yaml file
rm -rf /ragflow/conf/service_conf.yaml
while IFS= read -r line || [[ -n "$line" ]]; do
# Use eval to interpret the variable with default values
eval "echo \"$line\"" >> /ragflow/conf/service_conf.yaml
done < /ragflow/conf/service_conf.yaml.template
/usr/sbin/nginx
export LD_LIBRARY_PATH=/usr/lib/x86_64-linux-gnu/
PY=python3
if [[ -z "$WS" || $WS -lt 1 ]]; then
WS=1
fi
function task_exe(){
JEMALLOC_PATH=$(pkg-config --variable=libdir jemalloc)/libjemalloc.so
while [ 1 -eq 1 ];do
LD_PRELOAD=$JEMALLOC_PATH $PY rag/svr/task_executor.py $1;
done
}
for ((i=0;i<WS;i++))
do
task_exe $i &
done
while [ 1 -eq 1 ];do
$PY api/ragflow_server.py
done
wait;