[mythtv-users] NFS and remote backend
rbsteffes at gmail.com
Mon Jan 14 19:37:38 UTC 2008
On Jan 14, 2008 1:20 PM, Patrick Ouellette <pat at flying-gecko.net> wrote:
> On Mon, Jan 14, 2008 at 01:05:15PM -0500, Patrick Ouellette wrote:
> > >
> > > >
> > > > Are you running two NICs in the master box (one on the gigabit lan and
> > > > one on the 100 mega bit lan) or are you relying on a combination switch
> > > > (10/100/1000) to handle the different bit rate conversion?
> > > >
> > > > Pat
> > > > --
> > >
> > > I'm running through a D-Link firewall/router to handle it.
> > So you are expecting an NFS client set for 8K blocks to work well on a
> > network with a server and another client set at 32K blocks. I wonder if
> > that has anything to do with it.
> I reading that on the list, it sounds a bit harsh. Not meant to be an
> attack, more thinking out loud.
> The other thought that comes to mind is are you optimizing your
> database? A clean database is a happy database.
I would only expect that to be a problem if network latency was
suffering across the board, but it isn't. If I was getting bad
network timings in general, I'd think adjusting that might fix it, but
the problem seems to be exclusively to reading the files off the NFS
share. I run the optimize script on the DB regularly, but I haven't
gone into tweaking mysql parameters. Based on what other people do, I
wouldn't think that /should/ be necessary. I'd be a lot happier
blaming network settings if it failed both ways, instead it seems to
be something about the nature of the NFS that's failing.
More information about the mythtv-users