volumes in progress...

Dockerfile

@@ -17,7 +17,8 @@ RUN apt-get update && apt-get install -y \
 
 RUN echo "UseDNS no" >> /etc/ssh/sshd_config && \
 	echo "PermitRootLogin yes" >> /etc/ssh/sshd_config && \
-	echo "PasswordAuthentication yes" >> /etc/ssh/sshd_config
+	echo "PasswordAuthentication yes" >> /etc/ssh/sshd_config && \
+	echo "StrictModes no" >> /etc/ssh/sshd_config
 
 RUN pip install -U pip
 
@@ -35,4 +36,7 @@ RUN pip install pillow
 
 COPY ray_on_golem /app/ray_on_golem/
 
+RUN rm -r /root/.cache
+RUN mv /root /root_copy
+
 VOLUME /root

examples/GPU.Dockerfile

@@ -17,7 +17,8 @@ RUN apt-get update && apt-get install -y \
 
 RUN echo "UseDNS no" >> /etc/ssh/sshd_config && \
 	echo "PermitRootLogin yes" >> /etc/ssh/sshd_config && \
-	echo "PasswordAuthentication yes" >> /etc/ssh/sshd_config
+	echo "PasswordAuthentication yes" >> /etc/ssh/sshd_config && \
+	echo "StrictModes no" >> /etc/ssh/sshd_config
 
 RUN wget -O miniconda.sh https://repo.anaconda.com/miniconda/Miniconda3-py310_23.11.0-2-Linux-x86_64.sh
 RUN bash miniconda.sh -b -u
@@ -41,3 +42,8 @@ RUN pip install numpy numba
 RUN pip config set global.index-url https://pypi.dev.golem.network/simple
 
 COPY ray_on_golem /app/ray_on_golem/
+
+RUN rm -r /root/.cache
+RUN mv /root /root_copy
+
+VOLUME /root

examples/dds-with-ray.yaml

@@ -15,7 +15,7 @@ provider:
     node_config:
       # Parameters for golem demands (same for head and workers)
       demand:
-        image_tag: "golem/ray-on-golem:0.9.0-py3.10.13-ray2.9.3"
+        image_tag: "blueshade/ray-on-golem:0.10.0-dev-240404-5-py3.10.13-ray2.9.3"
 
         # List of urls which will be added to the Computation Manifest
         # Requires protocol to be defined in all URLs

examples/mandelbrot.yaml

@@ -24,7 +24,7 @@ provider:
       demand:
         # if not provided, image_tag will be autodetected based on currently used python and ray versions
         # check available versions at https://registry.golem.network/explore/golem/ray-on-golem
-        image_tag: "blueshade/ray-on-golem:0.10.0-dev-240402-1-py3.10.13-ray2.9.3"
+        image_tag: "blueshade/ray-on-golem:0.10.0-dev-240404-5-py3.10.13-ray2.9.3"
 
         # List of urls which will be added to the Computation Manifest
         # Requires protocol to be defined in all URLs

examples/outbound.yaml

@@ -28,7 +28,7 @@ provider:
       # Parameters for golem demands (same for head and workers)
       demand:
         # Check available versions at https://registry.golem.network/explore/golem/ray-on-golem
-        image_tag: "blueshade/ray-on-golem:0.10.0-dev-240402-1-py3.10.13-ray2.9.3"
+        image_tag: "blueshade/ray-on-golem:0.10.0-dev-240404-5-py3.10.13-ray2.9.3"
 
         # List of urls which will be added to the Computation Manifest
         # Requires protocol to be defined in all URLs

examples/outbound_tester.yaml

@@ -26,7 +26,7 @@ provider:
       # Parameters for golem demands (same for head and workers)
       demand:
         # Check available versions at https://registry.golem.network/explore/golem/ray-on-golem
-        image_tag: "blueshade/ray-on-golem:0.10.0-dev-240402-1-py3.10.13-ray2.9.3"
+        image_tag: "blueshade/ray-on-golem:0.10.0-dev-240404-5-py3.10.13-ray2.9.3"
 
         # List of urls which will be added to the Computation Manifest
         # Requires protocol to be defined in all URLs

golem-cluster.full.yaml

@@ -34,7 +34,7 @@ provider:
       # Parameters for golem demands (same for head and workers)
       demand:
         # Check available versions at https://registry.golem.network/explore/golem/ray-on-golem
-        image_tag: "golem/ray-on-golem:0.8.0-py3.10.13-ray2.9.2"
+        image_tag: "blueshade/ray-on-golem:0.10.0-dev-240404-5-py3.10.13-ray2.9.3"
 
         # List of urls which will be added to the Computation Manifest
         # Requires protocol to be defined in all URLs

golem-cluster.mini.yaml

@@ -16,7 +16,7 @@ provider:
     node_config:
       # Parameters for golem demands (same for head and workers)
       demand:
-        image_tag: "golem/ray-on-golem:0.9.0-py3.10.13-ray2.9.3"
+        image_tag: "blueshade/ray-on-golem:0.10.0-dev-240404-5-py3.10.13-ray2.9.3"
 
 # Tells the autoscaler the allowed node types and the resources they provide
 available_node_types:

golem-cluster.override.2-image.yaml

@@ -2,4 +2,4 @@ provider:
   parameters:
     node_config:
       demand:
-        image_tag: "blueshade/raytest:0.10.0-dev-py3.10.13-ray2.9.3"
+        image_tag: "blueshade/ray-on-golem:0.10.0-dev-240404-5-py3.10.13-ray2.9.3"

golem-cluster.yaml

@@ -33,7 +33,7 @@ provider:
       # Parameters for golem demands (same for head and workers)
       demand:
         # Check available versions at https://registry.golem.network/explore/golem/ray-on-golem
-        image_tag: "blueshade/raytest:0.10.0-dev-py3.10.13-ray2.9.3"
+        image_tag: "blueshade/ray-on-golem:0.10.0-dev-240404-5-py3.10.13-ray2.9.3"
 
         # List of urls which will be added to the Computation Manifest
         # Requires protocol to be defined in all URLs

ray_on_golem/server/services/golem/golem.py

@@ -291,16 +291,22 @@ async def _upload_node_configuration(
         ip: str,
         ssh_public_key_data: str,
     ):
+        async def run_command(cmd: str):
+            result = await context.run(cmd)
+            await result.wait()
+            logger.debug("Command executed: %s: %s",cmd, [e.to_dict() for e in result.events])
+
         provider_desc = await self.get_provider_desc(context.activity)
         logger.info(f"Running initial commands on {provider_desc}, {ip=}, {context.activity=}")
         hostname = ip.replace(".", "-")
-        await context.run("echo 'ON_GOLEM_NETWORK=1' >> /etc/environment")
-        await context.run(f"echo 'NODE_IP={ip}' >> /etc/environment")
-        await context.run(f"hostname '{hostname}'")
-        await context.run(f"echo '{hostname}' > /etc/hostname")
-        await context.run(f"echo '{ip} {hostname}' >> /etc/hosts")
-        await context.run("mkdir -p /root/.ssh")
-        await context.run(f'echo "{ssh_public_key_data}" >> /root/.ssh/authorized_keys')
+        await run_command("echo 'ON_GOLEM_NETWORK=1' >> /etc/environment")
+        await run_command(f"echo 'NODE_IP={ip}' >> /etc/environment")
+        await run_command(f"hostname '{hostname}'")
+        await run_command(f"echo '{hostname}' > /etc/hostname")
+        await run_command(f"echo '{ip} {hostname}' >> /etc/hosts")
+        await run_command("mv /root_copy/.bashrc /root_copy/.profile /root 2> /dev/null")
+        await run_command("mkdir -p /root/.ssh")
+        await run_command(f'echo "{ssh_public_key_data}" >> /root/.ssh/authorized_keys')
 
     async def _start_ssh_server(self, context: WorkContext, ip: str):
         provider_desc = await self.get_provider_desc(context.activity)
@@ -313,7 +319,7 @@ async def _restart_ssh_server(self, context: WorkContext, ip: str):
         try:
             await context.run("service ssh restart", timeout=120)
         except Exception:
-            msg = f"Failed to restart SSH server {provider_desc}, {ip=}, {context.activity=}"
+            msg = f"Failed to restart the SSH server {provider_desc}, {ip=}, {context.activity=}"
             logger.warning(msg)
             logger.debug(msg, exc_info=True)
         else: