Skip to content

Instantly share code, notes, and snippets.

View xxrjun's full-sized avatar

Hsiang-Yu Tsou xxrjun

  • National Taiwan University
  • Taipei, Taiwan
  • LinkedIn in/leon-tsou
View GitHub Profile
@xxrjun
xxrjun / .multi_ssh_helpers
Created October 15, 2025 03:27
Multi SSH Helpers
#!/usr/bin/env bash
# original source: https://lambda.ai/blog/how-to-serve-deepseek-r1-v3-on-gh200
# runbootstrap: setup known_hosts and authorized_keys (for passwordless ssh, you need to have your public key in ~/.ssh/id_rsa.pub or ~/.ssh/id_ed25519.pub)
# runip: run command on a specific ip (set ip=host:port or user@host:port)
# runk: run command on the k-th ip in the ips file
# runhead: run command on the first ip in the ips file
# runips: run command on multiple ips in parallel (set ips="..." or read from ips_file)
# runall: run command on all ips in the ips file
#!/bin/bash
# verify_mnnvl_health.sh
# Local verification for MNNVL compute tray & NVLink switch health.
# - Runs all checks to completion; no early exit.
# - Safe output mode ON by default (redacts IP/MAC/long HEX/IDs).
# - Prints a final Summary and returns nonzero if any check failed.
#
# Reference: https://docs.nvidia.com/multi-node-nvlink-systems/mnnvl-user-guide/verifying.html
@xxrjun
xxrjun / lvm-note.md
Last active November 14, 2023 15:21
了解LVM

LVM (Logical Volume Manager)

紀錄自己學習LVM

LVM 的優點

  • 方便調整 Logical Volume 的大小
  • 可以安裝任何 Filesystem 在 Logical Volume 上
  • 易餘替代故障的 Disk
@xxrjun
xxrjun / extend-lvm.md
Last active May 28, 2025 04:07
了解LVM並於Ubuntu擴充伺服器磁碟空間