Re: [BackupPC-users] backuppc slow rsync speeds
2012-09-19 13:19:51
<backuppc AT kosowsky DOT org> wrote on 09/18/2012 09:51:11
PM:
> Timothy J Massey wrote at about 12:54:35 -0400 on Monday, September
17, 2012:
> > I have several very similar configurations. Here's
an example:
> >
> > Atom D510 (1.66GHz x 2 Cores)
> > 4GB RAM
> > CentOS 6 64-bit
> > 4 x 2TB Seagate SATA drives in RAID-6 configuration
> > I get almost 200 MB/s transfer
rate from this array...
> > 2 x Intel e1000 NICs in bonded mode.
> >
>
> I snipped out most of Tim's original post because it seems that nobody
> has referenced the fact that he is using pretty low powered
> chips. Certainly, I could understand how compression would be slow
on
> these chips and could be the rate limiting steps.
You are correct. I've said from the beginning
that this is *embedded* class hardware. I'm trying to figure out
exactly what I can-or have to--do to make this work.
Again, I know that I can throw hardware at this; I
don't want to. I want to find what I can do to tweak my settings
for an acceptable level of performance. If I can't, then I will address
the hardware, and be confident that I understand why I am.
> Tom's Hardware benchmarked the Atom D510 against
circa year 2000
> Pentium 4 single core processors and found that for non-multithreaded
> programs that don't take advantage of new instruction set
> enhancements, they are pretty equivalent.
Ouch. I guess I bought into the Intel hype too
much... :(
I wasn't expecting insane performance. But I
admit I was expecting Core-type performance rather than Pentium 4! It
sounds like I'm actually getting Pentium III performance, though. (Which
makes sense: the Pentium M was a development of the Pentium III,
and it wouldn't surprise me if the Atom--designed for ultramobile--had
more in common with the Pentium M than Core...)
> So, I wouldn't be surprised if the problem is
using a "netbook"
> category performance processor to perform a computationally intensive
> server-type job...
Nor would I. At all. But throwing hardware
at it is not really answering the question, it's more addressing the symptom.
Besides, from someone who has run BackupPC on a wall
wart, I would have expected more interest in the challenge! :)
From my limited testing, I was able to get 4 x performance
simply by disabling compression. That takes my 4-day fulls and makes
them 1-day. If I can possibly double the performance again, I get
down to 12 hours. Mission accomplished. So, I'll keep tweaking.
As a bit of a sanity check for me, the motherboard
I'm using (SuperMicro X7SPA-HO-F) has an Atom D510 and is $175. A
SuperMicro X9SPV-LN4F-3LE with 3rd Gen Core i7-3612QE is $800. The
lowest-end X9SPV with an i3 is $600. Given that the target price
for the entire device is $1200 or so, that's a tough fit... Which
is why I'm trying to make the inexpensive, compact, cool and quiet Atom-based
boards work.
Timothy J. Massey
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________
BackupPC-users mailing list
BackupPC-users AT lists.sourceforge DOT net
List: https://lists.sourceforge.net/lists/listinfo/backuppc-users
Wiki: http://backuppc.wiki.sourceforge.net
Project: http://backuppc.sourceforge.net/
|
|
|