nvidia-smi hangs indefinitely after ~66 days

(github.com)

97 points | by tosh 2 hours ago ago

13 comments

  • wincy an hour ago

    Crazy, so if I understand correctly, something with B200s and nvlink is causing issues where after 66 days and 12 hours of uptime, nvidia-smi and other jobs start failing, timing out, then once you restart the cluster it starts working again.

    They suspect jobs will work if you only use 1 B200, but one person power cycled so wasn’t able to test it. Hopefully they won’t have to wait another 66 days for further troubleshooting.

    • layla5alive an hour ago

      Some 32-bit counter somewhere used when in NVLINK overflows?

      • mook 36 minutes ago

        Isn't 32bit counter 49 days? Assuming that one was counting milliseconds, at least.

        Only remember that because that's the limit for Windows 95…

        • repiret 9 minutes ago

          100ns intervals. My favorite part of that story is how long after Windows 95 was released before anybody discovered the bug.

  • yoshicoder 15 minutes ago

    I wonder if the process to debugging this is just to search for what power of 2 times a time unit equals ~66 days

  • userbinator 11 minutes ago

    I think it's an overflow of a scaled counter.

    Also, who else immediately noticed the AI-generated comment?

  • nulone 34 minutes ago

    NVLink postRxDetLinkMask errors show up right before the hang. Has anyone captured a bug report or stack trace while nvidia-smi is stuck to see what it's blocking on?

  • blackoil 42 minutes ago

    *China specific code leaked into mainline.

  • grayhatter an hour ago

    a pet peeve of mine, (along with people brigading on issues/threads e.g. posting them to unrelated news sites... op....) is woefully incorrect language.

    > at day 66 all our jobs started randomly failing

    if there's a definable pattern, you can call it unpredictabily, but you can't call it randomly.

    • toast0 5 minutes ago

      IMHO, what they said means that on day 65 all jobs work, on day 66, jobs work or don't, seemingly at random.

      But what they seem to be indicating is that all jobs fail on day 66. There's no randomness in evidence.

    • paulddraper an hour ago

      Unexpectedly is probably what they meant

    • JohnLeitch 14 minutes ago

      Seems quite predictable given the others in the bug report encountering the same.

  • zeehio 34 minutes ago

    66 days 14 hours and 24 minutes (66.6 days) would have been a far more diabolical hang...