[OpenAFS] multiple afs client, or have two caches

Rainer Toebbicke rtb@pclella.cern.ch
Thu, 11 May 2006 14:52:52 +0200


Todd M. Lewis wrote:

> If you've got a list of files that you really want to keep cached, just 
> put the file names into a list and do something like this:
> 
> while (1) ; do
>    for f in `cat ~/.afs.cache.list` ; do
>        head $f > /dev/null
>        sleep 10
>    done
>    sleep 600
> done
> 

as AFS does not cache files but chunks, the above solution would just 
keep the first chunk in the cache, not the complete file.

Also, Ted did not say that files were only used for read. If yes,
the need for a RAM cache is relative: on modern systems, if files are 
read frequently then they will reside in the system's buffer cache anyway.

I agree though that being able to pin chunks in the cache could have 
its merits, for example on a notebook if you know that the next time 
you connect it'll be over a slow line. Can't be that big a change...

-- 
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Rainer Toebbicke
European Laboratory for Particle Physics(CERN) - Geneva, Switzerland
Phone: +41 22 767 8985       Fax: +41 22 767 7155