infiniflow
[Question]: Can't start ragflow-es-01 as it keeps restarting
Describe your problem
m4 mackbook pro docker desktop Set the vm.max_map_count using this: docker run --rm --privileged --pid=host alpine sysctl -w vm.max_map_count=262144 set the ulimit -c unlimited
2025-02-06 10:33:55 # A fatal error has been detected by the Java Runtime Environment: 2025-02-06 10:33:55 # 2025-02-06 10:33:55 # SIGILL (0x4) at pc=0x0000ffff77e67b1c, pid=8, tid=17 2025-02-06 10:33:55 # 2025-02-06 10:33:55 # JRE version: (21.0.1+12) (build ) 2025-02-06 10:33:55 # Java VM: OpenJDK 64-Bit Server VM (21.0.1+12-29, mixed mode, sharing, tiered, compressed oops, compressed class ptrs, serial gc, linux-aarch64) 2025-02-06 10:33:55 # Problematic frame: 2025-02-06 10:33:55 # j java.lang.System.registerNatives()V+0 [email protected] 2025-02-06 10:33:55 # 2025-02-06 10:33:55 # No core dump will be written. Core dumps have been disabled. To enable core dumping, try "ulimit -c unlimited" before starting Java again 2025-02-06 10:33:55 # 2025-02-06 10:33:55 # An error report file with more information is saved as: 2025-02-06 10:33:55 # /usr/share/elasticsearch/hs_err_pid8.log 2025-02-06 10:33:55 [0.009s][warning][os] Loading hsdis library failed 2025-02-06 10:33:55 # 2025-02-06 10:33:55 # The crash happened outside the Java Virtual Machine in native code. 2025-02-06 10:33:55 # See problematic frame for where to report the bug. 2025-02-06 10:33:55 # 2025-02-06 10:33:32 /usr/local/bin/docker-entrypoint.sh: line 43: 8 Aborted ( elasticsearch-keystore create ) 2025-02-06 10:33:35 /usr/local/bin/docker-entrypoint.sh: line 43: 8 Aborted ( elasticsearch-keystore create ) 2025-02-06 10:33:42 /usr/local/bin/docker-entrypoint.sh: line 43: 8 Aborted ( elasticsearch-keystore create ) 2025-02-06 10:33:55 /usr/local/bin/docker-entrypoint.sh: line 43: 8 Aborted ( elasticsearch-keystore create )
Here for your information.
Thank you for the information. Do I need to install Elasticsearch directly on macOS instead of using Docker?
Hello, have you fixed this problem? I met with the same problem that ragflow-es-01 keeps restarting. Could you give me some instructions on fixing it? Thank you
MEM_LIMIT=26073741824
yes, I increased the MEM_LIMIT to 26073741824, but the problem still exists. The ragflow-es-01 logs still says that " A fatal error has been detected by the Java Runtime Environment:"
The same problem on M4 mackbook pro , docker desktop. Have you fixed this problem? Could you give me some instructions on fixing it? Thank you
I found one solution to this specific problem occurring in a M4 chip Mac. The answer is quite simple. JUST put two lines of codes in the docker-compose-base.yml
Like this,
services: es01: container_name: ragflow-es-01 ... environment: ... - "ES_JAVA_OPTS=-XX:UseSVE=0" - "CLI_JAVA_OPTS=-XX:UseSVE=0" ...
I found one solution to this specific problem occurring in a M4 chip Mac. The answer is quite simple. JUST put two lines of codes in the docker-compose-base.yml
Like this,
services: es01: container_name: ragflow-es-01 ... environment: ... - "ES_JAVA_OPTS=-XX:UseSVE=0" - "CLI_JAVA_OPTS=-XX:UseSVE=0" ...
Thank you! It works! However, I am unable to parse PDF files. The task is queued but gets stuck at the beginning.
I found one solution to this specific problem occurring in a M4 chip Mac. The answer is quite simple. JUST put two lines of codes in the docker-compose-base.yml
Like this,
services: es01: container_name: ragflow-es-01 ... environment: ... - "ES_JAVA_OPTS=-XX:UseSVE=0" - "CLI_JAVA_OPTS=-XX:UseSVE=0" ...
same as SkywalkerZH, worked but not able to parse pdf files
Patience, guys. It usually takes me half an hour to handle a 10 page pdf file.
Since I increased the upper limit of the container's memory limit, the problem has been resolved, and the phenomenon of ragflow-es-01 restarting repeatedly has never occurred again. Of course, apart from that, I haven't made any other changes to the configuration.
（Spec：M1 Pro 32G, 10 + 16 cores）
Patience, guys. It usually takes me half an hour to handle a 10 page pdf file.
Just increase the tokens of block, it works.