On Fri, Nov 03, 2000, Joe Cooper wrote:
> > Will commit the code to 2.4 in a few days unless someone has any reason
> > not to.
>
> I'm not going to argue that it shouldn't go in because this is much more
> stable than the current 2.3 async code (which is probably unusable), but
> I think there are some stability issues still needing to be addressed.
> Now that the cache is filled, and I've tried running the rest of the
> benchmarks, the box seems to regularly crash pretty hard. I'm going to
> try it on a ext2 partition next to see if the problem is with ReiserFS
> (the box is freezing due to a ReiserFS deadlock condition...but I think
> it is triggered by a Squid issue, since I don't see this problem with
> 2.2STABLE5+hno--I don't think...I've never run this set of benchmarks on
> that Squid before).
>
> Anyway...Here's what happens:
Good! Someone else sees a memory hog condition with aufs! I'm not blind.
Gotta figure out exactly why that is now ..
ADrian
Received on Fri Nov 03 2000 - 04:18:21 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:12:54 MST