• Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 513
  • Last Modified:

SQL Backup times in new ESXi cluster

I'm not sure if SQL backups would be a good way to gauge performance, but this warrants some concern.

We recently finished a project where we stood up a new virtual environment. Our old system was ESX 4 with a Clariion SAN serving FC LUNs. The FC connections were rated at 4GBs.

The new system is ESXi 5 with a NetApp NAS serving NFS datastores (IP storage is 4x 1GiGe).

As we migrated all of our applications and utility servers, we noticed a performance increase with the response times. We left our file server and main SQL server for last.

Our SQL(2008 R2) backups were taking about an hour each night to run. This is SQL's native backup process. We expected this time to change once we moved the VM over to the new system. It hasn't. If anything, it increased by a few minutes. With the new hardware on the new hosts we expected to see some decrease in time.

Am I missing something here?

Old ESX Specs:
HP DL380 G7
X5650 2.67GHz CPUs

New ESX specs:
Cisco C260 M2
E7-2870 2.4GHz CPUs

Memory contention was never an issue here.
  • 4
  • 3
  • 2
2 Solutions
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
4GB Fibre Channel to NFS.

I suspect NFS could be the bottleneck and slower.

Have you enabled Jumbo Frames. Test with and without Jumbo Frames enabled.

HOW TO: Enable Jumbo Frames on a VMware vSphere Hypervisor (ESXi 5.0) host server using the VMware vSphere Client

and see this recent Closed Question, on a similar NFS High Latency performance issue.

MMRNLAITAuthor Commented:
our average read latency is 4.51, write is 1.5 Milliseconds. When looking at the history of the datastore log, we see a spike of 28/67 milliseconds during the backup window.

I've attached a screenshot.

Looking at this graph, the numbers look good until we hit our backup window (10PM). Would'nt this be normal given the amount of I/O that is involved during a SQL backup?

Is enabling jumbo frames going to make that much of a difference? Reason I am asking is this will require an outage on the network side all the way to the core.

Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
Jumbo Frames can make an improvement, but it could make it worse.

It's usually a recommended option when use NFS.

Enabling jumbo frames, needs to be be enabled, on the ESX server, network switch and NetApp filer. So it's not a quick change, and will require planning before implementation.
Easily Design & Build Your Next Website

Squarespace’s all-in-one platform gives you everything you need to express yourself creatively online, whether it is with a domain, website, or online store. Get started with your free trial today, and when ready, take 10% off your first purchase with offer code 'EXPERTS'.

You may suffer from fragmentation. This usually shows when running backups.

You can check this by running the command

wafl scan measure_layout volumename

This will tell you if you need to defrag the LUN. A one time defrag can be done with a command like this:

reallocate start -f [-p] pathname

This can also be scheduled on a regular basis, see the man page for the reallocate command.
MMRNLAITAuthor Commented:
Robocat, when I attempt this command from a SSH session on the filer it comes back with "wafl not found". This is 8.0.2P4 7-Mode.

Any ideas?
Try if it works in advanced mode

priv set advanced
wafl ...
priv set

Or you could try this command instead:

reallocate measure volname
MMRNLAITAuthor Commented:
Quick question Hanccocka, when you said
"Jumbo Frames can make an improvement, but it could make it worse."

How could this possibly make things worse? We are discussing this change inhouse now..

Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
We have tried Jumbo Frames, on many installations, and although recommended, it really is a matter of test after enabling, and checking it performance is better or worse.

for some of our clients, performance was no better, and some worse.

It depends on your network infrastrcucture, switches, Im told that some switches do not have adequate buffers, although do support Jumbo Frames.

So, test, test,testband test again, and see if JF works for your site.
MMRNLAITAuthor Commented:
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

Join & Write a Comment

Featured Post

Free Tool: ZipGrep

ZipGrep is a utility that can list and search zip (.war, .ear, .jar, etc) archives for text patterns, without the need to extract the archive's contents.

One of a set of tools we're offering as a way to say thank you for being a part of the community.

  • 4
  • 3
  • 2
Tackle projects and never again get stuck behind a technical roadblock.
Join Now