Files
SnarfCode/scripts/check_dcgm_status.py

22 lines
812 B
Python

import paramiko
import time
c = paramiko.SSHClient()
c.set_missing_host_key_policy(paramiko.AutoAddPolicy())
c.connect('140.44.4.90', username='rancher', key_filename=r'C:\Users\Snarf\.ssh\id_ed25519_snarfnet', timeout=10)
def run(cmd):
stdin, stdout, stderr = c.exec_command(f"echo Katzir476! | sudo -S sh -c 'export KUBECONFIG=/etc/rancher/rke2/rke2.yaml; PATH=$PATH:/var/lib/rancher/rke2/bin; {cmd}'")
out = stdout.read().decode().strip()
return out
time.sleep(5)
print("Pods status:")
print(run("kubectl get pods -n ai-stack -l app.kubernetes.io/name=dcgm-exporter -o wide"))
print("Logs from one pod:")
pod = run("kubectl get pod -n ai-stack -l app.kubernetes.io/name=dcgm-exporter -o jsonpath='{.items[0].metadata.name}'")
print(run(f"kubectl logs -n ai-stack {pod} --tail=20"))
c.close()