You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@tooling.apache.org by gs...@apache.org on 2026/08/17 10:03:17 UTC

(tooling-llmao) branch main updated: Provision should not run servers; config them.

This is an automated email from the ASF dual-hosted git repository.

gstein pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/tooling-llmao.git


The following commit(s) were added to refs/heads/main by this push:
     new 732e61f  Provision should not run servers; config them.
732e61f is described below

commit 732e61f14d55744f1f685f5720d05a82db938937
Author: Greg Stein <gs...@gmail.com>
AuthorDate: Mon Aug 17 05:01:14 2026 -0500

    Provision should not run servers; config them.
    
    The launcher should not run as part of provisioning. Instead, use the
    container's supervisor to run the vLLM launcher process. This also
    let's the provisioning process complete, recording such in the
    container and workspace.
    
    Co-authored-by: Claude Sonnet 5
---
 hosting/vast/provision.sh | 21 +++++++++++++++------
 1 file changed, 15 insertions(+), 6 deletions(-)

diff --git a/hosting/vast/provision.sh b/hosting/vast/provision.sh
index f058727..540ac4b 100644
--- a/hosting/vast/provision.sh
+++ b/hosting/vast/provision.sh
@@ -7,15 +7,24 @@ set -x
 mkdir -p /workspace  # doesn't exist initially; don't fail if it does
 cd /workspace
 
-# fetch launcher if needed
 curl -fsSL -o /workspace/launcher.py \
   https://raw.githubusercontent.com/apache/tooling-llmao/refs/heads/main/hosting/launcher.py
-chmod +x /workspace/launcher.py
 
-# fetch servers.yaml from asfquart
-curl -fsS -H "Authorization: Bearer $FLEET_KEY" \
+curl -fsSk -H "Authorization: Bearer $FLEET_KEY" \
   "$ASFQUART_URL/vllm/config/$VLLM_SET" \
   -o /workspace/servers.yaml
+  # -k for the self-signed cert — fine as a stopgap, but worth trusting the CA properly later
 
-python /workspace/launcher.py
-# or: nohup python ... &   if on-start must return
+cat > /etc/supervisor/conf.d/vllm-launcher.conf <<'EOF'
+[program:vllm-launcher]
+command=/bin/bash -c "source /venv/main/bin/activate && python /workspace/launcher.py"
+directory=/workspace
+autostart=true
+autorestart=true
+startretries=3
+stdout_logfile=/var/log/vllm-launcher.log
+stderr_logfile=/var/log/vllm-launcher.log
+EOF
+
+supervisorctl reread
+supervisorctl update


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@tooling.apache.org
For additional commands, e-mail: commits-help@tooling.apache.org