I have three VERY different machines that run AMAZINGLY stable with
linux-2.0.9 compiled with gcc-2.7.2.3 OR gcc-2.8.1.
One is a DNS, bootp, dhpc, sendmail alias server that has only
modest disk activity and it's all IDE. Old slackware 3.4 system
with the latest glibc1. enough memory NEVER swaps, 80MB, pentium,
IBM Aptiva
The second is a webserver, with atalk, and samba, with a mix of IDE
and SCSI drives. Old slackware 3.4 system with the latest glibc1 AND
glibc2.0.7. enough memory NEVER swaps, 132MB, pentium-pro, IBM
Aptiva
The third is a usenet news server with 5 BIG SCSI drives that get a
full USENET feed. suse 6.0 STOCK. serious lack of memory, 32 MB of
ram usually has about 16MB swapped, pentium, IBM Aptiva.
These will ALL run for months without ANY intervention with 2.0.9.
With kernels 2.2.11 with Alan's EMERGENCY patches,
http://www.linux.org.uk/VERSION/relnotes.2211.html, with gcc 2.7.2.3
or with kernels 2.2.12, gcc-2.7.2.3, processes DIE with no log
messages.
The usenet machine will NEVER get through a skim. The skim dies with
a BUS ERROR. ...innd usually dies once a day for NO reason. The DNS
machine and the news server both have the crond and syslogd die with
no reason given.
As an added bonus the DNS machine upon shutdown gives a PANIC with no
register dump.
I have no problem running 2.0.9 forever, really, but I want to warn
anyone running anything newer that you are definitely running a risk
if you do so.
-- Jeffrey Hundstad Minnesota State University, Mankato http://www.mankato.msus.edu/jeffrey/
- 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/