Bug#401358: marked as done (System freezes under heavy disk IO)
Your message dated Mon, 11 Dec 2006 19:21:07 +0100
with message-id <20061211191219.5A1E.PETER.HIRDINA@gmx.net>
and subject line Bug#401358: System freezes under heavy disk IO)
has caused the attached Bug report to be marked as done.
This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.
(NB: If you are a system administrator and have no idea what I am
talking about this indicates a serious mail system misconfiguration
somewhere. Please contact me immediately.)
Debian bug tracking system administrator
(administrator, Debian Bugs database)
--- Begin Message ---
Package: linux-image-2.6.18-3-k7
Version: 2.6.18-6
After installing the version 2.6.18-6 of the kernel, my machine (SMP,
AthlonMP) just freezes under heavy disk IO (not immediately, but after a
few minutes).
I tried the modules dpt_i2o and i2o_block for my Adaptec 2400A raid
controller, and even tried JFS instead of XFS as filesystem, but it didn't
make any difference. Well, the only difference was the time it took until
the machine froze (actually with JFS it happened faster than with XFS).
I went back to the version 2.6.17-9 and the machine seems to run stable
again, even with more IO than needed to crash 2.6.18-6. So, it looks like
somehwere between those versions, a SMB related bug was introduced in a
more generic place, since the crashes happened across drivers and
filesystems.
regards,
Peter
---------------------------------------------------------------------
Peter Hirdina Frauenstrasse 128
Peter.Hirdina@gmx.net 89073 Ulm / Germany
---------------------------------------------------------------------
--- End Message ---
--- Begin Message ---
Version: 2.6.18-6
Sorry, my fault. There is no bug in the kernel.
Doing some further testing in this area, I noticed that APIC isn't
properly implemented on that board. Booting a kernel with "nosmp" just
spills out a lot of APIC errors, but boots just fine with "nosmp noapic".
Therefor I tried the "noapic" parameter also for SMP mode. And surprise,
the system suddenly runs rock solid under heady disk IO.
So, it was no code problem, just a hardware issue, and that could be
solved by using the kernel parameter "noapic".
cheers,
Peter
---------------------------------------------------------------------
Peter Hirdina Frauenstrasse 128
Peter.Hirdina@gmx.net 89073 Ulm / Germany
---------------------------------------------------------------------
--- End Message ---
Reply to: