Anthony's Blog: Using System Storage - An Aussie Storage Blog
anthonyv 2000004B9K 7,209 Views
In case you read my previous post and are wondering why it appears to finish early, I appear to have done some funky HTML editing and broken the post.
If you are particularly keen, check out the full (working) blog entry here:
This will be my last blog post for the year as I am taking a break for Christmas and New Years. I really want to thank you all for reading and following my blog.
2011 has been an amazing year with some great highs and some terrible lows. I certainly hope that 2012 is a bright year for everyone in the world, especially those whose lives have been affected by natural disaster and political upheaval.
To close, I want to wish you all a great holiday season, a Merry Christmas, a happy Hanukkah and a Happy New year and share with you three of the most important messages to take into the holiday season:
Firstly, stay safe. Here is a brilliantly put together (and totally serious) road safety message from New Zealand:
Secondly, stay healthy. If you can dance like these guys in a clip from the classic 1973 film of Jesus Christ Superstar, you will be very fit indeed! It is like a Zumba class on steroids (the video has a slow start, it really kicks off at 0:37 but it is worth the wait).
Thirdly, lets hope in 2012 that all of the worlds leaders will focus on peace and love. Here is a 1980s power ballad (yet again I show my age) with a very Christmas themed video clip. Enjoy.
I am getting this question on a very regular basis:
"We have just upgraded to ESXi 5.0 but we cannot find the VAAI driver on the IBM Website"
The answer? There is no vendor supplied driver because no driver is needed. ESXi 5.0 uses a SCSI T10 compliant set of commands that all vendors need to support for VAAI to work.
But of course in the tradition of all answered questions, it leads to another question:
"Once I have upgraded to ESXi 5.0 how can I tell if VAAI is really working?"
The good news is that it is very easy to spot if ESXi 5.0 has detected a VAAI capable LUN. The moment a new LUN is detected by ESXi 5.0 it tries out an Atomic Test and Set command. If that works, you will see that Hardware Acceleration shows as Supported in vCenter. In the screen capture below I have three datastores, two from XIV and one from Storwize V7000, all presented to an ESXi 5.0 server. I dragged the Hardware Acceleration column over from the right hand side to help with the screen capture (in case your vCenter looks different), but you can see the Hardware Acceleration column shows each DataStore as Supported (and did so the moment the volume was detected).
Of course having seen the Hardware Acceleration Supported message only proves that Atomic Test and Set works. To confirm if XCopy (Hardware Accelerated Move) is working, on SVC or Storwize V7000 we can use the Performance monitoring panel. In the example below I first performed a storage vMotion, moving a virtual machine between two Datastores located on the same Storwize V7000 (running 22.214.171.124 firmware). I then performed a clone of the same virtual machine, where the source was on one datastore and the target was placed on another (but both located on the same Storwize V7000). What you can clearly see is that both operations (storage vMotion and cloning) generated no volume traffic, only MDisk traffic. This means that the ESXi server is doing none of the work and the storage is doing all of the work.
For more examples of how to test VAAI, check out section 9.5.5 Testing VAAI, in the new XIV IBM Redbook draft, XIV Storage System: Host Attachment and Interoperabilty which you can find here:
It documents the following tests:
The tests can be performed regardless of whether you have an XIV (on code levels 10.2.4a and above) or a Storwize V7000/SVC (on code levels 126.96.36.199 and above).
If upgrading to ESXi 5.0 with IBM Storage, please also be aware of the following knowledge base articles regarding VAAI support with IBM Storage:
The Storwize V7000 and SVC have a command line interface that you access via SSH. Every-time you logon, whether it is to transfer a file (using a tool like pscp), issue a single shot command from a script (using a tool like plink) or logon to issue commands interactively (using a tool like PuTTY), you clearly need to authenticate yourself. Since June 2003, the way you did this was to use a public/private key pair, where the SVC or Storwize V7000 had the public key and the SSH client (such as PuTTY) authenticated using the private key (the PPK file).
However with release 6.3 of the SVC and Storwize V7000 firmware, the use of key files is now optional. A user can now authenticate purely by using a password. This includes using your domain ID. So if you defined LDAP to your machine, as I documented here, you could now SSH direct to your Storwize V7000 or SVC, use your Domain user id and password and not go through the key file setup task. Nice!
The choice to continue to authenticate just with an SSH key remains available. If a user has both a password and a configured key file, then either method will work (you only need to use one - not both). Existing scripts will be unaffected by this change, so nothing gets broken because of this.
I think this is a very positive change and one I openly welcome. Combined with LDAP, this really makes user account setup an easy and simple task.
anthonyv 2000004B9K Tags:  512 wwpns 4096 2048 aix v7000 storwize svc npiv 3 Comments 15,292 Views
IBM recently released a new version of firmware for the SAN Volume Controller and Storwize V7000. This is known as release 6.3 and continues the tradition of two major updates per year, each adding significant new functions.
Since I blithely listed this feature in a recent post I have received lots of emails asking exactly what it means, so I thought I had better explain.
Most of these numbers are very high and few customers actually approach these maximums. The main issue I am seeing for some of our larger AIX customers is this one:
The reason this can become an issue is the combination of NPIV and AIX Live Partition Mobility. NPIV allows one physical HBA to be shared among multiple operating system instances, each one believing it has exclusive access to the HBA with each one allocated its own unique WWPNs. Suddenly a single HBA which used to present just one WWPN through the SAN to the SVC, can now present vast numbers of them. In addition AIX Live Partition Mobility (which lets you move AIX operating systems between LPARs on the fly) needs additional pre-configured WWPNs defined on the target LPAR to support the move. This further increases the quantity of WWPNs that need to be defined to the SVC (one easy way to spot NPIV generated WWPNs is they normally start with the letter C).
So the bottom line is that IBM needs to make this limit bigger and SVC and Storwize V7000 6.3 code contains the necessary architectural changes to allow this. The first phase is to start potentially supporting up to 2048 WWPNs per I/O group although clearly based on the initial version of the release notes, the long-term plan is to support 4096.
But there is a problem and it has nothing to do with the SVC or Storwize V7000. The problem is that there are certain SAN configurations which may have issues with these large numbers of WWPNs (mainly around older SAN switches not having the CPU power for the switch fibre channel name-server and login-server to handle vast numbers of WWPNs coming out of one HBA).
So what should you do if you need to push the limits?
Contact your IBM Pre-Sales support and ask for a SCORE request to be opened (also known as an RPQ). You will need to detail your current SAN configuration (especially switch models and firmware levels) so that SVC development can ensure you won't overwhelm your switches. It will also allow our development team to learn how many clients our there need this support. All approvals will include a requirement to upgrade to release 6.3, so you should include this in your planning.
Any questions? Feel free to leave a comment or send me a tweet or an email.
anthonyv 2000004B9K 5,027 Views
If you want to read something visionary, I just finished The Medium is the Massage by Marshall McLuhan. With dramatic changes occurring in the world, often driven by social media, it is amazing to read a book that understood this so well yet was written in 1967. In fact Marshall passed away in 1980, well before most people got near a TCP/IP address and when the role of the Internet in societal change was still many many years away. A fascinating read (with some amazing art work!).
I have also been reading the excellent Steve Jobs bio by Walter Isaacson, what a fantastic book! I did not know about the friendship between Steve Jobs and Larry Ellison, nor about Al Gores role in Apple. The details of the early collaboration between Steve Jobs and Bill Gates are really interesting, as is the amazing back story of how Pixar came to be. The book is as much a history of the personal computer as it is a bio of Jobs himself. Reading the book was also quite thought-provoking, especially on themes like: motivation; ignoring the impossible; creating A-teams and the impact of office layout on creativity and teamwork.
At one point the book refers to the Macintosh software dating game, which led me to watch this video. It is also a revelation. Enjoy!
The XIV GUI (which you can download here) is available for a very wide variety of platforms:
My gut feel is that the truly vast number of installations and downloads are for the Windows GUI version. I suspect the rest rarely get a look in, I mean do people actually use the XIV GUI on AIX, Solaris or HP-UX hosts? I really doubt it.
However you can also get just the XIV command line interface (also called the XCLI) for the following operating systems:
Now I can see why these would be popular. Being able to script XIV CLI commands and execute them locally makes perfect sense, so all the major Operating Systems are represented. But it is curious that a separate Windows CLI installer is not listed. Of course you get the XIV CLI when you install the Windows GUI, but I am equally curious if there are users who want to avoid having the GUI present on servers that only need the CLI.
An example of this would be if you use Commvault SnapProtect with XIV, since each client will need to issue XCLI commands to drive the hardware based snapshots. So the good news is that you can actually force the XIV GUI installer to install only the CLI component. You can do this by using the following command (in a command prompt with the XIV GUI installer file in that directory):
xivgui-3.0.1-build3-win.exe /SP- /NOCANCEL /VERYSILENT /NORESTART /TYPE="xcli" /LOG="%SYSTEMDRIVE%\xcli.txt"
You will need to change the file name at the start of the command to suit the version you have downloaded. I tested it on version 3.01 and it worked just fine, all it installed was the XIV CLI. So keep this in your back pocket and use when required.
And if you ARE using the XIV GUI on AIX, HP-UX or Solaris, I would love to hear which platform you are using and why. And if you are still using Windows ME.... your persistence is admirable.
anthonyv 2000004B9K Tags:  response mbps performance svc v7000 times storwize iops 11 Comments 20,701 Views
With the 6.3 release of the Storwize V7000 and SVC code (which I blogged about here), there are so many new features and functions that I have plenty more to blog about!
The first new feature I blogged about was LDAP support, but an existing feature that has been enhanced is the performance monitor (brought in with release 6.2). When this first came out I put a video on You Tube showing what metrics could be displayed in that release. This is a sped up image with no voiceover:
Now with release 6.3 IBM has added separate graphs for reads and writes plus the ability to display IOPS or MBPS, plus the ability to display graphs of read and write latency. Nice! I got so excited I made another You Tube video, this one with narration. So now you can compare the new to the old:
Now this is interesting: IBM is offering a ratings system that allows customers who bought IBM products to write reviews and leave ratings (out of five) on IBM Storage, Power and System Z products, straight from the main ibm.com website.
To get to the Ratings and Reviews page go here: http://www.ibm.com/systems/info/review.html
It's a really fast and easy way to give feedback and let other customers see it, so why not try it out!
anthonyv 2000004B9K Tags:  fibre simplify san zoning ds800 switch alias v7000 xiv storwize svc channel 2 Comments 23,553 Views
Lets imagine a new rack server or a new blade server has been added to your Fibre Channel SAN. The first job for the SAN administrator is to zone it to the storage it requires access to. The task normally runs something like this:
The main trap here is that when creating a zone, you need to ensure you select all of the correct storage aliases for your selected storage device. For instance we could have a simple layout like this:
Fabric 1 contains our new server (in this example an IBM x3850) and three XIV ports:
This means when creating the zone I need to identify and select four separate aliases. What I could do instead is create an alias with all my XIV target ports in it. Now I only have two aliases to select in that fabric:
If this was a Storwize V7000 implementation I could do the same thing. A typical install often look like this, where fabric 1 contains our new server and two Storwize V7000 ports:
This means when creating the zone I need to identify and select three separate aliases. What I could do instead is create an alias with both my Storwize V7000 WWPNs in it. Now I only have two aliases to select in that fabric:
This method of amalgamating multiple storage port aliases works fine for devices like DS8000, SVC, Storwize V7000 and XIV. I use this method all the time to simplify zoning and I find it reduces both mistakes and the time required to complete zoning tasks.
The only exceptions are:
I would love to hear any techniques you have to make your (and my) life easier.