Re: NTFS Mounting does not work reliably on current versions of NetBSD

[Michael van Elst <mlelstv%serpens.de@localhost>]

On Mon, Sep 01, 2014 at 10:57:37AM -0400, William D. Jones wrote:

> If that is the case, how should I debug my issues further then? Nothing
> TCP/IP intensive works properly, and yet I know the hardware is capable of
> it from running Transmission, ntfs-3g and friends on Raspbian. Is this a
> kernel buffer management issue then?


Maybe this was a bit misleading. I just wanted to say that you don't have to
rebuild the kernel with a different NMBCLUSTERS value but can use the sysctl
command to increase (but not decrease) the number of (and /etc/sysctl.conf to
make it persistent).

Network buffers are allocated from two pools:

% vmstat -m | egrep '^Name|^mbpl|^mclpl'
Name        Size Requests Fail Releases Pgreq Pgrel Npage Hiwat Minpg Maxpg Idle
mbpl         256     9681    0     9524    88    82     6    15     1   inf    1
mclpl       2048     3996    0     3869   432   398    34    47     2  3000    2


mbpl are small buffers, currently 9681-9524=157 are allocated, using 6 memory 
pages,
there is no limit (Maxpg = inf).

mclpl are the larger cluster buffers, 3996-3869=127 are allocated, using 34 
memory pages,
there is a limit of 3000 pages (12000 buffers, 24MByte).

The Hiwat column will tell you how much had been allocated in the past.


Greetings,
-- 
                                Michael van Elst
Internet: mlelstv%serpens.de@localhost
                                "A potential Snark may lurk in every tree."