Over the last couple of months my team has been working hard with our partner and OEM, Hitachi Data Systems to help get native support for the AMS 500 in the operating system. MPxIO, our core path state and encapsulation technology with responsibility for high availability on a wide range of storage arrays now just works with the AMS as of build 84 of Open Solaris. This is a first because Hitachi was active in the development of the source and they contributed to the Open Solaris community. Additionally, we now cover nearly all mid-range and high end arrays in the Open Solaris operating system. Choice is something our customer demand.

I was at a high end retailer the other day and they were pleading for mid range support for the Hitachi arrays. They went on to articulate that every other bit of highly available storage just worked in Solaris and the only gap was mid-range Hitachi.

Give it a whirl if you are inclined. Yell at me with bugs. Need another array under MPxIO? Yell at me and we will guide you through the SCA, show you sample code and help you with the code review. The larger the community, the better for all.

 

Last week my HPC team posted the iSER project on opensolaris.org. Very cool and my first project that is developed from eggs to apples, totally in the open. Let me be clear, COMSTAR and various other projects are open sourced and in the open, but the iSER project is the first project I have that has an external open solaris coder working jointly with us. Qlayer sent Bart to work with us in December and Qlayer is now an active member of our development team. We are still working some of the source code repository issues, but, we are a true open source development team.

We have a little announcement about it on StorageStop, so drop in if you get the chance. If you want to help out, sign up for it over in open solaris. Or, contact me directly.

Anyway, well done Qlayer, well done Peter and team. Going to be a fun year.
 

I am playing a lot of hockey right now, trying to make up for all the missed drills and games due to my travel schedule. Last night I was practicing blocking the pass while playing defense. This consists of getting really low, sliding one leg on the ice and placing your stick across the passing lane. Best if you can do that at speed. I was doing a great job at that, if I do say so myself, right until I looked up and smashed face first into the boards. I was so focused on my form that I forget to check what was going on around me. Good life lesson... Got to keep your head in the game (shameless nod towards my kids infatuation with the latest incarnation of High School Musical) but, you also have to keep your head up and your eye on what is coming.

I played with the latest version of SDX and installed it on a vanillia x86 machine I have at home. Smooth all the way through. We continue to make progress and this is a good thing.

Happy 2008 to one and all. I will be more regular about posting. Perhaps nothing profound, but certainly some random thoughts.

I work with a lot of technologies. These are the current projects I am tracking:

http://opensolaris.org/os/project/comstar/
http://opensolaris.org/os/project/caiman/
http://opensolaris.org/os/project/crossbow/
http://opensolaris.org/os/project/crypto/
http://opensolaris.org/os/project/emulex/
http://opensolaris.org/os/project/honeycomb/
http://opensolaris.org/os/project/kmf/
http://opensolaris.org/os/project/mpxio/
http://opensolaris.org/os/project/npiv/
http://opensolaris.org/os/project/nws/
http://opensolaris.org/os/project/osd/
http://opensolaris.org/os/project/samqfs/
http://opensolaris.org/os/community/hpcdev/
http://opensolaris.org/os/project/smf-doc/
http://opensolaris.org/os/project/iscsitgt/
http://opensolaris.org/os/project/avs/
http://opensolaris.org/os/project/isns/

A note about the title:

I am going to "borrow" a post from David Weibel, a former Sun engineer. He writes well and he is also an excellent engineer. You can also borrow Sun equipment. Hopefully it will lead to a nice marriage. And, of course, I borrow the catchphrase of a world class New Orleans chef. Onwards...

I just got off a phone call with a vendor trying to implement high-availability on a Solaris client using their iSCSI target. I was gratified to hear how painful it was in other operating systems (and, by extension, painless in Solaris). Here at Sun, we try to make it simple to use the storage stack. Conform to the standard, and it should just work. Particularly if the device is standards-based. Asymmetric, roll your own, hand-crafted devices are great for tinkering, but if you want a world class storage subsystem, you should probably use a standards-based product.

 Here is a link to David Weibel's blog entry on iSCSI and MPxIO, showing how simple it is to hook up a standards based array.

Now, as Emeril Lagasse might say, "Let's kick it up a notch." BAM.  Imagine you want to build that storage device using Solaris as your storage operating system on the array. Hey, you can do that too, now that the Availability Suite, the iSCSI Target, and ZFS are all part of the operating system. Check it out on OpenSolaris.org. The world moves on. Value moves up the stack. A world class operating system with data services, RAID functionality and some of the best plumbing in the industry. Now all you need is a Thumper and a couple of X4600s to run it on.

Did you know you can try out a Thumper, an X4600 and AVS free for 60 days? Try it. Buy if you like it. Send me a note or a comment if you want to try this product set.

Also, 'cause I dig good food and good writing, check out this link. The author writes the review so well I gain virtual weight every time I read about the food she eats. 

I had about a dozen interesting openings I wanted to implement for this blog. However, turns out none of them were all that  interesting. It all felt forced. So, here is a cool blog. The author, Ben Rockwood, talks about iSCSI Target mode and Solaris. 

Worth a read.
 

Saturday night and I am writing a bit of code (yes, I still write code) and I feel compelled to comment. I am trying to use a seemingly simple function UuidtoString(...) and having a heck of a time of it. I keep getting a failure to convert to type error. These are common errors and it usually involves just figuring out the right way to hold your head when you enter the function. Anyway, after about an hour of scratching my frontal lobe, it suddenly occurs to me that the default project setting may have been "helping me out".

 It was, and I feel a bit foolish. But, that does bring me to my thought for the day. It occurs to me that this is a simple mistake to make, and had someone clever decided that the function help should include common mistakes, well, I would be a bit further along and a bit closer to watching some hockey with my wife. My point? Make things easy. There are no style points for keeping secrets when it comes to code.