Re: [BUG] How to crash 4.9.2 x86_64: vmscan: shrink_slab

[Sami Farin]

On Tue, Jan 10, 2017 at 10:22:41 +0100, Michal Hocko wrote:
> On Mon 09-01-17 23:02:10, Sami Farin wrote:
> > # sysctl vm.vfs_cache_pressure=-100
> > 
> > kernel: vmscan: shrink_slab: super_cache_scan+0x0/0x1a0 negative objects to delete nr=-6640827866535449472
> > kernel: vmscan: shrink_slab: super_cache_scan+0x0/0x1a0 negative objects to delete nr=-6640827866535450112
...
> > 
> > 
> > Alternatively,
> > # sysctl vm.vfs_cache_pressure=10000000
> 
> Both values are insane and admins do not do insane things to their
> machines, do they?

Not on purpose, unless they are insane :)

Docs say:
"Increasing vfs_cache_pressure significantly beyond 100 may have
negative performance impact."
Nothing about crashing.

But anyways, the problem I originally had was:
with vm.vfs_cache_pressure=0 , dentry/inode caches are reclaimed
at a very alarming rate, and when I e.g. rescan quodlibet media
directory (only 30000 files), that takes a lot of time..  I only download
some files for a minute and dentry/inode caches are reclaimed,
or so it seems.  Still, SReclaimable keeps on increasing, when it gets to
about 6 GB , I increase vm.vfs_cache_pressure .... 

-- 
Do what you love because life is too short for anything else.
https://samifar.in/