On Sat, 8 Jan 2000, David S. Miller wrote:
>
> Date: Sat, 8 Jan 2000 12:06:15 -0800 (PST)
> From: Robert Dinse <nanook@eskimo.com>
>
> 2.2.14 Appears to be more stable on my SS-10's than 2.2.13 was but I
> did get a spin_lock crash this morning.
>
> ...
>
> The 4/670 that I have with same CPU's was stable under 2.2.13 and
> continues to be stable under 2.2.14.
>
> Are these identical kernel images?
No, they are different. The 4/670MP has three SCSI controllers, one of
which is the PTI so it has both drivers configured in, and it has RAID
parititons, so that is configured in.
> Based upon that, the next question is, are your compilers identical
> on both systems?
>
> Later,
> David S. Miller
> davem@redhat.com
Compiler on the 4/670MP is:
Reading specs from /usr/lib/gcc-lib/sparc-sun-linux/egcs-2.90.27/specs
gcc version egcs-2.90.27 980315 (egcs-1.0.2 release)
Compiler on the SS-10 is:
Reading specs from /usr/lib/gcc-lib/sparc-unknown-linux-gnu/egcs-2.91.66/specs
gcc version egcs-2.91.66 19990314/Linux (egcs-1.1.2 release)
Note that this version was tried at someone's suggestion (I think Anton
suggested it) earlier.
I have also built 2.2.13 kernels on the 4/670MP and then ran them on the
SS-10, and they too are not stable on the SS-10.
An extra piece of data that may be worth while, the SS-10 is a web server
running Apache 1.3.3p3, the 4/670MP is a news server running INN 2.2.2.
The reason why I mention this is that I also have a number of LX's running
SparcLinux; they are all stable except one that is running Apache 1.3.3p3. The
others all do other things.
The LX that runs apache, processes, things like rpc.mountd, rpc.nfsd,
portmapper, inetd, etc, randomly die IF I allow Apache to run more than 50
clients, if I keep MaxClients below that they keep running.
Aside from Apache, I have LX's being IRC servers, a mail server, and
running IRC services, and they are all stable. I also have one running an SSL
server, and it is stable, but it gets so few hits that it's never more than
about ten clients running.
I don't know if this is relevant or not, but these Hypersparc modules are
the type that are asynchronous to the bus; the faster modules they put out
later were synchronous (some multiple of bus speed), but these are not.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sat Jan 15 2000 - 21:00:13 EST