[mythtv-users] How do I get coverart/etc in watch recordings?

[Jean-Yves Avenard]

2009/9/17  <f-myth-users at media.mit.edu>:
> That's good to hear.  Presumably, the NFS parameters in my test don't
> exactly match yours, and no doubt you're running a newer kernel.  For

2.6.28 kernel

> the sake of the archives (since "the wiki" might change in the future
> and someone reading your message can't possibly know what those values
> were right now), can you relate the exact /etc/fstab and /etc/exports
> values you were using on each end, and your kernel version?  Thanks!

Using exactly the option showed on the wiki
/etc/fstab
192.168.10.11:/data     /data        nfs
rsize=8192,wsize=8192,hard,intr,nfsvers=3,actimeo=0 0 0

on the server:
/etc/exports
/data           (rw,no_root_squash,insecure)


>
> And you're sure this was options on client, and client push, right?
> It might also be interesting to see if noac is different from actimeo=0.

Quite frankly, I don't care much ... I just want to set it up and
forget about it

>
> (For my tests, I could see the change in behavior just by doing
>  mount remount -o actimeo=1 server:/foo /foo
> on the client while pushing to the server; changing options
> mid-transfer and looking at the rate of growth of the file showed I
> could nail or fix performance by a factor of 50 just by flipping back
> and forth on the mount as the transfer was in progress.)

Certainly didn't notice this at all...

with actimeo=0, transfer of data via nfs for 60s:
42060528.86 bytes/sec

(I have people watching liveTV right now, so it's impacting the speed)

with actime=1

37746628.46 bytes/sec

So go figure , it's now faster !

That just confirms that is makes no difference on my install

>
> At some point, I may see if I can narrow down some pathological set
> of NFS options that cause noac/actimeo=0 to behave so badly for me;
> that won't be today, though.

good luck