Yet another apology post about going silent on my blog… sorry guys, life has just been crazy.
So, let’s catch everyone up, and then get into current projects. As of November, 2012, my role with my company changed from an Enterprise Architect/SAN Architect/Linux Architect to the Manager of the Infrastructure and Operations division within IT. So I now lead up a team of 16 awesome individuals within my company. It’s been a long 6+ months for me, trying to adapt to my new role, and working to optimize many areas within our I&O team.
I attended the Gartner ITxpo Symposium in Orlando in late 2012, which was nothing short of awesome. I’ve been to several conferences, and as far as IT Strategy and Management, this was bar none the best conference I’ve ever been to.
Some of our current projects include evaluating TEMS (Telecom Expense Management) providers to offload all Telecom Expense Management and Billing, which will result in a huge cost savings for us. If your company is not doing this now, I highly suggest talking to a couple different vendors to determine what opportunities exist for you.
We are doing a lot of work around a complete redesign of our Architecture strategies going forward, so my team has been diligently evaluating hundreds of different intersections of software, hardware, and services to find a best scenario for us. It has been a lot of fun, and we are hopefully winding down on finalizing our I&O strategy and roadmap for the next couple years.
As some of these projects move forward, look for updates from me on any technologies or services we’ve evaluated, and my thoughts.
I apologize for the long time since my last blog post, life has taken over and made me quite busy. When I’m not working, I’m trying to maximize time with the family, and that leaves little time for anything else.
We’re now 6+ months into our 3PAR arrays in production, with no major issues, and some amazing performance numbers. The Cisco Nexus switches have been great, but Cisco apparently has a major issue with using san port channels over their ONS DWDM product, so much so that you cannot do it. A bug has been entered, but we have to wait until the next release of NX-OS for this to get resolved.
On top of all that, we’re beginning discovery for implementing a fresh new ERP, which means my role as an Architect has come into full effect, and I’ve been quite a busy man.
Shoot me any questions or comments, maybe even some more topics you’d like to see ( but please no Microsoft products, I’m very lacking in my MS skillset )
We left off with a Linux host that had been built and configured to begin the installation of 3PAR System Reporter. Let’s continue on and get System Reporter installed and configured.
As a recap from Part 1, we are building a System Reporter system in Linux. I have chosen to use RedHat Enterprise Linux 5.8, along with MySQL for the database backend. Both System Reporter and MySQL will be hosted on the same server. Read More…
Installed an Emulex CN1000E CNA PCIe card into a test HP Proliant DL360 G7 today, so we could play with configuring and managing a CNA card on our new Cisco Nexus 5596UP switches. We connected everything together with 3M Cisco Passive Twinax (DAC) cables, and ran VMware Vsphere 5.0 on the server.
First step is to cable everything up, then log into the Cisco Nexus switch, and setup FCoE, as well as configure the ports. Read More…
System Reporter is a 3PAR tool that provides informative reporting about your 3PAR arrays, as well as being able to build custom reports, and schedule daily execution and email of any report.
System Reporter is also required to be installed, in order to use Adaptive Optimization, as System Reporter is used to determine which chunklets are to be migrated between which CPG’s. All the Adaptive Optimization configuration is done from within System Reporter’s web interface.
System Reporter can be installed on Windows 2003, Windows 2008, or RedHat Enterprise Linux 5. It may use either Microsoft SQL, MySQL, Oracle, or SQLite as it’s backend database, although there are restrictions on each database. MySQL seems to be the recommended database to use, as it has the fastest query times due to it’s MyISAM structure, and it has minimal restrictions across the server Operating Systems.
I will be installing and configuring System Reporter 2.8 on a RedHat Enterprise Linux 5.8 virtual machine. I will be using MySQL as the database backend, which will be installed on the same host virtual machine. In part 2 of this series, I will be then be installing and configuring System Reporter. Part 3 will be configuring and adding a 3PAR InServ system, in my case a T400. Part 4 will be setting up Adaptive Optimization. Part 5 will be all about reports, and how to create custom and scheduled reports. Read More…
We started up and configured two 3PAR T400’s this past week, one with 160TB usable storage, and the other with 100TB usable storage. Each have around 2.5TB of SSD in them, the rest are 600GB 15K rpm FC disks.
They each arrived as two seperate racks, with the controller nodes and cages (shelves) installed, but the disk magazines were seperate, in crates and other boxes. Overall each one was quite a few boxes to have to unpack and organize for installation.
I’ll try to break down the components a bit, so it will be easier to follow along with the installation. 3PAR has some unique names for their components. Check out their Concepts Guide for some even more detailed information.
Each of our T400’s came with 2 racks, the 1st rack containing the Service Processor (SP), 4 Controller Nodes, and 6 cages. A cage is what 3PAR calls their disk shelves. Each cage holds 10 Magazines, and each Magazine holds 4 disks, for a total of 40 disks per Cage. The second rack only contained 2 cages. Each cage is connected to the controllers using stress-free OM3 Fiber. The stress-free cabling is VERY important, as the cage covers will not safely attach unless you use stress-free cabling, available from HP or 3PAR.
Once everything has been unpacked and unboxed, the array frames are rolled into place. They do not bolt together, they simply sit next to each other. Once all underfloor cabling required has been run, the rubber feet are ratcheted into place, making sure the wheels are just slightly off the floor, and the array is level.
Installation of the cage magazines comes next. Each magazine has 4 drives in it front to back, and definitely have some weight to them, if they are FC or NL drives. They simply slide into the cages, and lock in with ease with a handle. There is also a screw that is screwed down to prevent the locking handle from coming loose. This process takes some time, and make sure you are wearing your handy dandy anti-static strap.
Once all the magazines have been inserted, your local HP field rep will run through his/her checklist of items, which includes verifying cabling, correct power hookup, all magazines are correctly inserted, etc. Once everything is verified, it’s time to throw the breakers on those Eaton PDU’s in each cabinet, and feel the huge gush of air come from the intense fans in the system. You may need to throw the mini breakers on each power supply on the rear of the T400 cages as well, along with the batteries. Make sure all lights are green.
The next steps are performed by the authorized HP installers, which is configuration using the SP (Service Processor), with a 3PAR Central engineer on call as well. This process can take several hours, as it’s an extensive verification process, along with a mini-stress test of an hour against all the hardware. There are a lot of components to configure, so just keep your HP installer hydrated and fed, and soon enough he/she will finish up.
Once the HP installer is done, you can fire up your InForm CLI or InForm management console, and start provisioning.
Overall, I was impressed with how throrough HP and 3PAR are with this array. We have multiple XP class arrays, as well as several EVA arrays from HP, and none of them came close to the amount of dedication for installation that the 3PAR array did. Kudos to 3PAR for setting such a high standard, and kudos to HP for keeping that through the acquisition.
Here’s a video after startup:
In the following days/weeks, I’ll be blogging a bit more about my experiences testing/poking/prodding the arrays, including some posts on replication, recovery manager, performance monitoring and tuning, and anything else I can come up with.