Re: Bug: slab corruption in 2.5.67-mm1

[J Sloan <joe@tmsusa.com>]

Joshua Kwan wrote:

>On Sat, Apr 12, 2003 at 03:05:22PM -0700, J Sloan wrote:
>  
>
>>I had run 2.5.67-mm1 for some days and for
>>the most part it ran well in it's duties as dns,
>>squid, vpn/firewall and postfix server, with
>>the only oddity being the ide messages which
>>I reported earlier.
>>    
>>
>
>You're insane running a -mm kernel on a production machine, IMHO.
>
Oh please.

This is my network, and there is nothing
insane about it - FYI the stock 2.4.18 RH
kernel had severe stability problems on
this system - OTOH 2.5.67-mm1 is like
the rock of gibraltar in comparison.

:-)

>They're good for desktop use but in my experience anything that needs to
>stay up for more than a few days should at LEAST use one of the
>stability-oriented patches like -mjb or -osdl, if not the vanilla
>kernel.
>
Sorry,  -mm fixes a number of bugs that
render the vanilla kernel unusable for me.

>2.5-mm typically goes nuts with such errors as you described after a
>few days of uptime, as far as I've seen and noticed. These bugs will
>probably eventually be fixed, but at this time, it's still unstable.
>
By running 2.5 and reporting, maybe we
can help expose those bugs - but since .65
or so, I haven't really seen any show stopper
issues - everything works, and works well,
on my particular hardware, and workload.

>
>I wouldn't run anything above 2.4.20 on a box that does what you
>describe..
>  
>
My clients get official vendor kernels, but
in the sanctity of my own domain I run the
latest and greatest and take a look at what's
coming - what I'm running today is a sneak
preview of what my clients will be running
tomorrow.

Joe