[mythtv-users] Raid Performance Tweaking

[Debabrata Banerjee]

RAID 5 sequential writes are not slower, they are almost as fast as a RAID 
0. RAID5 random writes are slow, but still not slower than a single disk 
unless you're running like 1 byte random IO's.

MythTV only does sequential writes. It will be blazingly fast regardless of 
what chunk size you use.

Napkin calculations 19.2Mbits * 2 + 8Mbits = 4 Megabytes per second. If you 
can't do this you have other problems Even if you suppose that RAID 5 writes 
are much slower than even a single disk.

I have an HD tuner, an MPEG 2 encoder, and a software encoder at 4-8Mbits on 
a single 5400 rpm disk and I'm not even close to having disk problems. I can 
even move files on and off the disk while encoding.. I've even done a pvmove 
to another disk while encoding just for kicks.. No issues.

You probably have other problems. PCI bus contention, for example. Very high 
CPU usage during HD playback, etc. Put the backend or frontend in another 
machine and see what happens.

Personally every encoder I have is in a separate PCI bus. The master backend 
NIC is in another PCI bus. None of the encoder chips play well with other 
PCI cards. You just end up with an unstable machine.

----- Original Message ----- 
From: "Dean Wilson" <dean.k.wilson at gmail.com>
To: "Discussion about mythtv" <mythtv-users at mythtv.org>
Sent: Thursday, July 06, 2006 5:54 PM
Subject: Re: [mythtv-users] Raid Performance Tweaking


In my case, the mythbox is also hosting a web and email server, and is
acting as a firewall for all the internal computers.  I also have two
high def, and one standard def tuners, which frequently all record at
the same time.  So there's plenty of disk read/writes, which seems to
be slowing playback down.

In using it with 5 disks configured with Raid5, (which, to my
understanding, is slower to write to than when writing to a single
disk, since it has to write additional parity bits to the nth drive) I
noticed that HD playback stuttered any time that it was also recording
another HD show.  I had DMA turned on, but haven't been able to get
past my playback issue when the drives were being written to.  I'm
currently reconfiguring my system anyway, so I thought I'd run the
test to see which configuration worked best.

(The rest of my system shouldn't be suspect, as it's an AMD X2 4200+
with 1GB ram.)

Any other configuration thought/suggestions would be welcome...
~Dean

On 7/6/06, Debabrata Banerjee <davatar at comcast.net> wrote:
> Whatever you do it's going to be irrelevant for mythtv unless you add a
> massive amount of tuners. Unless you decide to make a RAID array out of 
> 4GB
> dis\ks from 1996.. Even one disk should be good for many streams of video.
> Let's not confuse Mbit and MBytes, as people often do.
>
> ----- Original Message -----
> From: "John Drescher" <drescherjm at gmail.com>
> To: "Discussion about mythtv" <mythtv-users at mythtv.org>
> Sent: Thursday, July 06, 2006 5:06 PM
> Subject: Re: [mythtv-users] Raid Performance Tweaking
>
>
> >I think the problem here is that guide is very outdated. Did you see the
> > computer they were using "Dual Pentium Pro 150 MHz"? I bet that was on
> > a 2.2kernel also so to me any numbers there are meaningless. At work I
> > use 256K
> > or 512K blocks on my 1TB or 2TB raid 5 or raid 6 arrays. My latest 
> > server
> > build featured 6 X 320 GB seagate 7200.10 drives in a raid 6 256K 
> > clusters
> > for ~ 1.2TB of raid space. With that setup on a Athlon64 M2 3200 system
> > (asus mobo) I get ~260 MB/s on the non cached hdparm benchmark.
> >
> > John
> >
> > On 7/6/06, Dean Wilson <dean.k.wilson at gmail.com> wrote:
> >>
> >> I know that many of us are interested in Raid, but don't really know
> >> the best configuration for a myth setup.  I, like many others here, am
> >> interested in Raid5, and would like it to perform optimally for (very)
> >> large files.
> >>
> >> I've read http://www.tldp.org/HOWTO/Software-RAID-HOWTO-9.html, which
> >> gets into chunk and block size performance testing, but doesn't use
> >> chunks above 32K.  I would think that a larger chunk size would be
> >> ideal for myth, as it uses exceedingly large files.
> >>
> >> I figured I'd use bonnie to measure performance gains/losses, since
> >> that's what the howto used.  Obviously I'll play with different
> >> chunk/block size combinations, and will report my findings.  But I'm
> >> curious to know if anyone out there would be interested in any other
> >> specific tests?  If so, let me know -- I'll try to include them, and
> >> will put together a web page with the results I find.
> >>
> >> Thanks,
> >> ~Dean
> >> _______________________________________________
> >> mythtv-users mailing list
> >> mythtv-users at mythtv.org
> >> http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users
> >>
> >
>
>
> --------------------------------------------------------------------------------
>
>
> _______________________________________________
> mythtv-users mailing list
> mythtv-users at mythtv.org
> http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users
>
> _______________________________________________
> mythtv-users mailing list
> mythtv-users at mythtv.org
> http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users
>
_______________________________________________
mythtv-users mailing list
mythtv-users at mythtv.org
http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users