[Asterisk-Users] RAID affecting X100P performance...

Mike Benoit ipso at snappymail.ca
Wed Jul 21 17:26:39 MST 2004


I didn't want to turn this in to a software vs. hardware raid, or IDE
vs. SCSI. I was more curious about the PCI bus/interrupt issues and the
mainboard. I only have 1 line going in to this asterisk server, so CPU
usage is not an issue whatsoever. Even during a raid rebuild.

Asterisk's CPU usage doesn't even show up on the charts.

zttest and the odd complaint is the only way to tell if the problem is
occurring. I've actually done some more testing since my post and
discovered the problem happens even with a SINGLE HD in the box. If I
work the drive hard enough, zttest starts reporting accuracy down to
80%. With two drives I can get it as low as 2%.

I did a quick test with a Celeron 2.4Ghz, and brand new mainboard and
couldn't seem to reproduce an accuracy drop below 99.98%.

So it seems like the X100P (possibly other Digium cards too?) is _very_
picky about the mainboard its used with. Not only have I had echo issues
that are specific to some mainboards, I've now also discovered "beeps"
or "cutting out" with some mainboards. The two don't seem to be linked
either. Since the mainboard that is experiencing "cutting out" is one
that has had the least amount of echo so far. However I do wonder if the
"cutting out" has a negative effect on the echo canceler or echo
training.

Really working two HD's on the mainboard (RAID or not) really brings the
problem to light, but the problem is even visible (though not near as
bad) with a single HD.


I would be interested in seeing if other people can reproduce low zttest
accuracy rates with their mainboards. zttest is in the zaptel/
directory, and you can run it while Asterisk happily chugs along
handling calls.

What I usually do is run zttest in one window, then in another window
run "updatedb" or "find / > /dev/null" or "hdparm -t /dev/hda" or all of
them at the same time. The harder and longer you work the drives the
better. If the problem exists on your system, you'll see the zttest
accuracy drop below 99.98% at some points.


On Wed, 2004-07-21 at 16:25 -0700, Kevin P. Fleming wrote:
> Scott Laird wrote:
> 
> > That hasn't been my experience at all.  Frankly, I've never seen a cheap 
> > (<$3k) hardware RAID controller that can touch software RAID's 
> > performance on Linux, especially in "challenging" setups, like RAID-5.  
> > Sure, software RAID eats more CPU, but most PCs have CPU to spare these 
> > days.  Would you rather eat 10% of one of your Xeon CPUs to get 200 
> > MB/sec or 100% of an Intel 960 to get 15 MB/sec?
> 
> While this is certainly true, in the context of Asterisk you also have 
> to consider the extra PCI bus usage for all this data going back and 
> forth to the drives while the RAID parity/mirror stuff is being done.
> 
> > In the context of Asterisk, where disk I/O is either logging or 
> > voicemail, buying a 3ware card and a pair of IDE drives seems like a 
> > decent business decision.
> 
> Yep, it works very well.
> _______________________________________________
> Asterisk-Users mailing list
> Asterisk-Users at lists.digium.com
> http://lists.digium.com/mailman/listinfo/asterisk-users
> To UNSUBSCRIBE or update options visit:
>    http://lists.digium.com/mailman/listinfo/asterisk-users
-- 
Mike Benoit <ipso at snappymail.ca>




More information about the asterisk-users mailing list