ZFS leaking vnodes (sort of)

[Brian Donnell]

I put these patches in on my i386, took out the i386 recommended tunings and
increased the max vnodes as recommended for amd64.  Last time I did this I
was able to cause kmem_map errors very quickly.  With these changes,
however, I'm not seeing any problems yet.  I'm running through as many tests
as I can think of for file accesses and transfers between and out of zfs
pools.  So far so good.  Other than the memory usage I haven't really
noticed a change in performance.  Note this is a pretty high spec'd machine
I'm testing on, though.

-- Brian

On 7/8/07, Pawel Jakub Dawidek <pjd at freebsd.org> wrote:
>
> Whoa! Nice catch... The patch works here - I did some pretty heavy
> tests, so please commit it ASAP.
>
> I also wonder if this can help with some of those 'kmem_map too small'
> panics. I was observing that ARC cannot reclaim memory and this may be
> because all vnodes and thus associated data are beeing held.
>
> To ZFS users having problems with performance and/or stability of ZFS:
> Can you test the patch and see if it helps?
>
>