apr-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "William A. Rowe, Jr." <wr...@rowe-clan.net>
Subject Re: Support for fstatfs
Date Mon, 30 Oct 2006 21:18:38 GMT
Graham Leggett wrote:
> William A. Rowe, Jr. wrote:
> 
>> Well, seeing as any test you do will be befuddled by file compression
>> on the
>> volume you are writing and that trusting this information is foolhardy
>> if there
>> are multiple users of the device, I'd ask you to rethink the design
>> before you
>> create yet another set of "My drive has XXX free but the program
>> crashes with
>> a No Space Left message!!!" bug reports :)
> 
> Trusting this information is obviously not foolproof, but using this to
> define a high water mark, beyond which caching quietly stops (with a
> logged explanation, of course) would be very useful, rather than rushing
> headlong into the wall, as is down now.
> 
> Call it a subtle and controlled degradation of service, if you will.

It seems that a more effective high-water mark is to set an upper bounds
on the desired cache size.

Mime
View raw message