Comments on: VMTN Storage Performance Thread and the EMC Celerra NS-120

By: jason

jason — Thu, 08 Sep 2011 13:04:18 +0000

Which link specifically? All 4 links in the first paragraph are working. I verified this morning.

Jas

By: Paul Aviles

Paul Aviles — Thu, 08 Sep 2011 05:15:38 +0000

Jason, the link seems to be broken. Can you recheck it please?

Thanks,

Paul

By: Welcome to vSphere-land! » Storage Links

Welcome to vSphere-land! » Storage Links — Fri, 08 Jul 2011 15:13:44 +0000

[...] – VMware vSphere and ESX 3.5 Multiprotocol Performance Comparison Using FC, iSCSI, and NFS (NetApp) VMTN Storage Performance Thread and the EMC Celerra NS-120 (Virtualization Evangelist) Comparing the I/O Performance of 2 or more Virtual Machines SSD, SATA [...]

By: jonathan

jonathan — Sat, 02 Apr 2011 06:35:07 +0000

Thanks. A great article which is clear and helpful. Actually, I also want to try to duplicate your performance result.

By: Jumbo Frames Comparison Testing with IP Storage and vMotion » boche.net – VMware Virtualization Evangelist

Mon, 24 Jan 2011 14:04:05 +0000

[...] [...]

By: jason

jason — Tue, 09 Nov 2010 19:47:18 +0000

@Paul:
I’ve uploaded the Iometer benchmark file and a description text file to the following location:
http://boche.net/dropbox/iometer/

By: Paul Aviles

Paul Aviles — Tue, 09 Nov 2010 15:32:19 +0000

Jason, can you post the IOmeter configuration files for the test you conducted? I am trying to duplicate your performance and I cannot get any way closer to your numbers.

Regards,

Paul

By: iSCSI or NFS with EMC Celerra ? « GOING VIRTUAL

iSCSI or NFS with EMC Celerra ? « GOING VIRTUAL — Wed, 07 Apr 2010 02:22:12 +0000

[...] [...]

By: jason

jason — Mon, 01 Feb 2010 23:41:44 +0000

Russ, slight improvement. New clone time of a 270GB VM is 45 minutes from and to NFS.

By: Russ

Russ — Mon, 01 Feb 2010 21:28:31 +0000

Hey Jason,

RE: “Cloning a 270GB VM on NFS takes nearly an hour.”

Has cloning speed improved with these updates?

By: Chad Sakac

Chad Sakac — Mon, 01 Feb 2010 00:04:53 +0000

Thanks Jason! NFS performance on Celerra is right inline with iSCSI in general (minor variances depending on IO type and other factors), and both can meet a lot of customer needs. In turn, with small block IO in the 4-64K range ( which tends to be IOps bound), they perform nearly identically to FC. FC does pull away with larger IO sizes (>64K) which tend to be bandwidth bound (MBps), unless one switches to 10GbE (which is supported across the EMC midrange portfolio). This is due to the single TCP session (for data) nature of the NFSv3 client PER NFS mount.

The tweaks are important. I would HIGHLY recommend anyone using Celerra NFS to look at the steps we sent over to Jason, which Scott Lowe has posted here:

http://blog.scottlowe.org/2010/01/31/emc-celerra-optimizations-for-vmware-on-nfs/

The Celerra for VMware techbook is also important reading for any EMC Celerra NFS customer using VMware (there are also CLARiiON and Symmetrix with VMware techbooks). These are all publicly and openly available (and orderable in hardcopy if you so desire).

One that is Celerra specific, and very important is the “uncached” filesystem parameter. This means that the Datamover doesn’t cache the IO. Based on the hardware architecture of the Celerra, write/read caching is still done, but it’s done on the block part of the Celerra (underneath the filesystem). The Celerra filesystem read/write caching is tuned for general purpose NAS (wide variety of files of all different sizes), in the NFS on VMware use case, the pattern is different – dominated by a relatively small number of very large files (VMDKs). This paramter (which applies only to the particular filesystem, not the Celerra as a whole) has a very significant effect on performance.

A vcenter plugin is coming very shortly that automates all the filesystem creation and management (along with many, many other very cool things) on EMC Celerra NFS. Come to VMware partner exchange!

By: jason

jason — Sun, 31 Jan 2010 17:27:44 +0000

Yep I plan on it

By: Lars Troen

Lars Troen — Sun, 31 Jan 2010 09:23:47 +0000

You should also update the VMTN thread with your new finding. Everyone is not reading this blog yet.

By: jason

jason — Sun, 31 Jan 2010 06:02:01 +0000

Performance numbers updated in the blog post. NFS is a lot closer now to swISCSI after a few tweaks from EMC.

By: Leif Hvidsten

Leif Hvidsten — Fri, 29 Jan 2010 17:06:10 +0000

While I’m not an EMC user and only have experience with NetApp, Jonathan’s post may be onto something. In the excellent multivendor post on NFS by Chad, Vaughn, and others, it is mentioned to:
1. Enable the uncached write mechanism for all file systems (30%+ improvement.)
2. Disable the prefetch read mechanism for file systems consisting of VMs with small random accesses patterns.

http://virtualgeek.typepad.com/virtual_geek/2009/06/a-multivendor-post-to-help-our-mutual-nfs-customers-using-vmware.html

You could also check that your design follows some their general best practices as well as EMC’s docs. Though, I would think when performance testing just 1 VM’s I/O, a load balanced link aggregation setup isn’t going to improve anything since the traffic would be over 1 TCP session to 1 datastore, unless you’re connected to more than 1 datastore and generating I/O concurrently to each. However, your spindle setup would be a huge factor.