[OpenAFS] AIX 5.2 large cache success report

[chas williams - CONTRACTOR]

In message <dkaghe$9rv$1@sea.gmane.org>,Joe Buehler writes:
>Curiously, while running a lot of "cat" commands to fill the cache and
>make sure the machine didn't crash, the reported cache usage (fs
>getcache) was never observed to be above about 8,500,000.
>...
>This makes me wonder whether it is safe to set cacheinfo to the size of
>the partition?  For initial tests I had set it to 90% of the partition

probably not.  you need to allow for modern fs behavior.  when a file
is deleted, the disk blocks are not immediately ready for reuse for
some filesystems.  the actual delete is handled by a background task in
the kernel.  this is the case with ext3 -- i suspect ibm's jfs is similar.