[mythtv-users] random livetv stalls

Michael T. Dean mtdean at thirdcontact.com
Fri Feb 7 02:37:23 UTC 2014


On 02/06/2014 09:03 AM, Michael Watson wrote:
> On 6/02/2014 6:43 PM, Jean-Yves Avenard wrote:
>> I was mounting NFS with the option: rsize=8192,wsize=8192,timeo=14,intr
> I ended up with mount options 
> "proto=tcp,retry=10,rsize=32768,wsize=32768,hard,intr".  Shares are 
> mounted by automount, using nfs4.   Using nfs4 and the tuning of wsize 
> and rsize solved all my problems.   automount is used so the drives 
> can enter powersave.
>
> I have 2x SBE/FE and 1x FE reading and writing via NFS to storage 
> located on mbe.  I run a gigabit network with 1x FE on wireless

AIUI, on (even semi-)modern kernels (like for the past 5+ years?), you 
shouldn't specify rsize or wsize at all.  Changing it from 8kB to 32kB 
would improve things, but removing it completely (as Jean-Yves 
suggested) should /significantly/ improve things.  The kernel can 
negotiate far larger/faster values dynamically, if you don't constrain it.

Mike


More information about the mythtv-users mailing list