Data-driven recommendations, please.

Crazy as it may seem, as a Technical Consultant I spend a good bit of time gathering data — performance data, utilization data, configuration data…the list goes on. We even had a specific couple hour training session last Friday in Greensboro on some of the tools we should be using (thanks to Matt Hensley for spearheading this). One good thing about being an EMC reseller is that there’s no lack of tools for this (much less the various VMware and Cisco tools). I won’t say we don’t get frustrated at the quirks occasionally but they’re good stuff overall (it’s amazing what some people can write in their “spare time”).

Here’s a sampling of what I look at regularly… (multiple emails today around these exact items actually)

  • SP Collects — this is an easy way for a customer to pull a compressed full config file from an EMC array, send it to me, and I can generate a nice spreadsheet telling me a ton of information about the array (model, firmware levels, RAID group disk layout/sizing/utliization, LUN details, connected hosts, protocols, etc.). Just a boatload of config info…also helps customers know more about the state of their storage.
  • NAR Data — this goes deep into EMC performance land….stuff like cache usage, IOps per disk/RAID Group/LUN (i.e. too high), latency/response time in milliseconds per LUN (goes straight to application level performance issues), nice one page heat map (see at a glance…red = bad is the quick version). Suffice it to say if you export everything you can create a 100+ page PPT presentation. We usually gather NAR data over the course of a week…one day is just too short for storage performance (and sometimes longer than a week depending on the array size and usage).
  • Backup Exec Database Analysis — given a database dump, a nice analysis tool + some manual work, I can see a lot of information quickly about your Backup Exec environment (backup sizes, errors, schedules, trends, etc.). Not only does this give a nice report we can go over (help understand potential or actual backup environment issues), it also helps provide very accurate sizing if you’re interested in Data Domain or Avamar (“Never backup the same data twice” etc.).
    • We have simliar tools for NetBackup, etc.
  • Capacity Planner — great tool for VMware resellers which helps analyze physical servers (or remaining physical environments) and provide lots of details around CPU/RAM/disk space/disk I/O/etc. This is another tool run over time so as to give extended performance details. Also a customer favorite as it turns out a very nice presentation at the end (some of it is a bit confusing but overall very good stuff).
  • RVTools — this is a personal favorite available at It uses vCenter APIs to hook directly into vCenter and present you tons of information in a tabular format (think Excel talking to vCenter almost). What’s even better for me as a TC is that everything you see in RVTools can be directly dumped to an Excel file. If you haven’t used this tool, I highly recommend it for day to day usage…go check it out.

Believe it or not, that’s just scratching the service….we have a whole “EMC Tools” VM with 60+ tools on it which we keep regularly updated (and that’s just the EMC tools). While I have a lot of background in storage and virtualization, learning all the specific tools we use at Varrow has been one of the challenges for me over the last several months….one I’m still working on actually. It is a challenge I’ve appreciated though because….

Why all this time gathering data? Well, not to go buzzword happy, but the real goal is to make data-driven recommendations (this is a bit of a pet phrase for me for a while actually)….just makes for better solutions and therefore better relationships. When making recommendations, I want to be able to back them up with hard data (configuration, trending, etc.) as much as possible — the results of almost everything you see above is presented back to the customer and usually left in some kind of hard copy or soft copy. While we do sometimes have to use industry averages, it’s always far better to use live customer data despite the extra work.

Now, does this always happen? (“I need a configuration by tomorrow” makes that hard.) Of course not…but it is definitely the goal and usually the reality.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s