On 24/08/24(Sat) 13:14, Hugh Graham wrote: > On Sat, Aug 24, 2024 at 09:31:43PM +0200, Martin Pieuchot wrote: > > On 24/08/24(Sat) 12:09, Hugh Graham wrote: > > > The machine that slowly received the ports tree crashed upon reboot, > > > and I have included the traces. It was running 7.5-stable at the > > > time. I will wait an hour or so, and then restart and repeat > > > against current. > > > > Thanks, could you elaborate how you trigger the corruptions? I'd like > > to try to reproduce them locally. > > To summarise, I first encountered the crash while trying to move > a large filesystem archive from an old machine that could only > manage roughly 1MB/s to a relatively new machine over ethernet. > I then reproduced the crash by setting media 10BaseT on the > receiving machine, and sending the archive from another new > machine. I haven't been able to reproduce the crash with either > local or fast network transfers, but also haven't tested this > exhaustively. > > Subsequently, I have reproduced the crash by setting media 10BaseT > on my target machine, and sending 25GB of ports and packages via > pax over netcat. Possibly loops of CVSROOT would work better than > the large packages. There is a dmesg from the target machine > deeper in this email thread.
Hugh, could you please send the exact commands that triggered the crash? All the details matter!
