Conversation

Notices

1. Embed this notice
   Chris Siebenmann (cks@mastodon.social)'s status on Friday, 29-Nov-2024 01:10:03 JST Chris Siebenmann

   A belated obvious reason: ZFS's habit of panicing on assert failures and printing almost no diagnostics to identify the surrounding circumstances of the panic (like, say, what pool and filesystem and so on involved) is more understandable when I remember that Solaris had kernel crash dumps. So in theory ZFS would panic, the crash dump would get saved, and you'd go in with a debugger and get all of that context.

   None of that works on ZFS on Linux. (And panics are bad in general.)

   • Haelwenn /элвэн/ :triskell: likes this.
   • Embed this notice
     penguin42 (penguin42@mastodon.org.uk)'s status on Friday, 29-Nov-2024 03:20:28 JST penguin42

     in reply to

     @cks You can do crash dumps on Linux; they set up a kexec into a small setup held in reserved RAM and it does the dump; you then use 'crash' to poke at the dump file later. The RHEL etc set normally do that by default.
     (But I'll agree filesystems panicing is generally bad)

     Haelwenn /элвэн/ :triskell: likes this.