Diagnosing a Linux Performance Regression

(automattic.com)

36 points | by program 9 hours ago ago

4 comments

  • zdw 6 hours ago

    I wonder if this is still a problem if nftables was used instead.

    nftables has been in mainline linux since 3.13 over a decade ago, and has been in the default in distros for a while (Debian 10 had it as default 6 years ago), but K8s support for it has lagged far behind, with it just recently out of beta in 1.33, and still not the default.

  • boguscoder an hour ago

    It almost reads as a cliffhanger, the regression was caused by commit that seemed to be related to fixing synchronization issue , so I assume just plain revert wasn’t an option

  • leetrout 7 hours ago

    (January 2025)

    Not something recent in the kernel.

    • emmelaich 4 hours ago

      I suspect using more than two year old (or even older) kernel versions in production is pretty common.

      Anyone have stats on this?