[OpenAFS-devel] still seeing iput errors on 2.2.19...

[Chaskiel M Grundman]

Excerpts from mail: 14-Apr-101 Re: [OpenAFS-devel] still s.. by Nathan
Neulinger@umr.edu 
>     2048 bytes  DV 21 refcnt -27651
> That's a volume for a heavily accessed cgi script. That one appears to
> be the opposite problem though.
It's not the opposite problem, it's just that the i_count value is being
interpreted as a signed short by cmdebug (see the definition of struct
AFSDBCacheEntry in src/fsint/common.xg)

> I'm doing what I can to instrument afs_dcache.c to do some cache
afs_dcache.c doesn't manipulate refcounts on vcache/inode objects. All
that stuff hapens in afs_vcache.c and osi_vnodeops.c

What kernel does this happen with? I seem to be having a different sort
of refcount problem on my 2.4.3 systems -- the machine runs out of free
vcaches because all/most of them stay in use. This ends up triggering a
"VLRU cycle" panic.