Networker

Re: [Networker] What does good LTO3 performance look like?

2008-02-19 10:16:28
Subject: Re: [Networker] What does good LTO3 performance look like?
From: Ian G Batten <ian.batten AT UK.FUJITSU DOT COM>
To: NETWORKER AT LISTSERV.TEMPLE DOT EDU
Date: Tue, 19 Feb 2008 14:36:21 +0000
On 19 Feb 08, at 1349, Fazil.Saiyed AT anixter DOT com wrote:


Hello,
If you were to eliminate hardware bottlenecks, i wonder if using compression directive might improve backup performance ?

That's an interesting point. We're using zfs compression, which is obviously uncompressed as it comes up from the filesystem, then runs uncompressed over the wire and is then compressed by the LTO3 drives. It might indeed be worth playing with doing the compression within networker and sending it to a non-compressing device.


Are you using AFTD atall, if not perhaps a VTL or AFTD may also be worth a try.

Most of our backup is indeed to AFTD, or is done with replication. What I'm sending to tape is ~20TB of replicated data: we replicate from our live fileserver to an offsite clone, and then back up the offsite clone back into an on-site tape library for long-term retention. So performance isn't a big deal: I've doing 200GB/hour, say 4TB/day allowing for tape changes and random factors, and I need to do 20TB/month. So I've got a safety factor of six.


Using Jumbo frames if your GigE supports would yield the best results.

Interesting: we've had very limited benefits from jumbo frames with modern switches and TCP stacks: ~5% at best. What's other peoples' experience been?

ian


To sign off this list, send email to listserv AT listserv.temple DOT edu and type 
"signoff networker" in the body of the email. Please write to networker-request 
AT listserv.temple DOT edu if you have any problems with this list. You can access the 
archives at http://listserv.temple.edu/archives/networker.html or
via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER