SSD is Something to See

Nigel Griffiths has another great demonstration video, this one a discussion of the recently announced SSD drives. I'll give you some highlights here, but watching the entire clip is well worth the 22 minutes.

Nigel mentions that these drives evenly spread data throughout the memory cells on the device, and all of the cells can be accessed at once. He also states that you shouldn't attempt to defragment an SSD because doing so unnecessarily consumes some of the finite number of writes and shortens the drive's lifespan.

These disks currently ship with 69 GB usable out of a 128-GB capacity. This spare capacity allows the disk to actively search for suspect memory cells, mark them as bad and move data away from them without losing any capacity, since you have only half of the possible disk space available for your use.  Compared to traditional drives, SSDs can access every memory cell at once, so rotational latency and seek times aren't issues. Nigel shows random I/O per second (IOPS) in the range of 28,000, as opposed to 125-200 IOPS with a traditional disk drive.

He also notes that SSDs use less power than a traditional disk. In his example he compares 330 watts for five SSDs vs. 7,700 watts in a single DS4700 system. Disk per disk, he says SSDs require about half of the power of traditional disk subsystems.

In the demo Nigel covers the different disk carriers, machines and expansion drawers that the drives can be used in, going from 2.5-inch small form factor disks for blades to 3.5-inch SAS carriers. He also displays various pictures--including some of a dismantled SSD--to provide a more complete visual perspective.

To take advantage of SSD, you must update your machine firmware and make sure that HMC, AIX and VIOS are at supported levels. In the demo Nigel connects an SDD directly to his LPAR, and then runs lsconf to show the disk in his machine. Then, after running cfgmgr, you see that the disk doesn't immediately appear. Nigel goes into the menus--DIAG > task selection > RAID array manager > IBM SAS disk array manager--to create an array candidate pdisk and format to 528-byte sectors.

Once you take these steps, you'll be able to select the solid state drive that will format the disk. At this point, you can see a pdisk, which can put into an array. From the same IBM SAS disk array manager menu, you can create an SAS disk array by selecting the controller, RAID level, stripe size and the pdisk you just created. You can then you can go back and list the array, which will show you the hdisk and pdisk mappings that you just created.

If this doesn't seem clear, watch the demo. It should make more sense once you see it performed on a live system.

Now when he runs lsconf, the hdisk appears, and it can be used as would any other disk -- you can add it to a volume group, create filesystems, etc.

Nigel runs some tests to demonstrate SSD speed. To prepare for testing, he unmounts his test /SSD filesystem and remounts it with the cio option. By running a tool called ndisk64, he performs some random I/O tests, runs iostat and can see 11,000 IOPS and 44 MB/sec.

For the second test, he uses the raw device and gets 27,000 IOPS to the disk. He then mentions the rumors he's heard about even faster times being recorded in the labs.

So why do we care about SSD? Customers with intense performance requirements can put heavily accessed data onto an SSD. Nigel also brings up using SSD in conjunction with active memory sharing and paging spaces, as SSD will provide much faster access than regular disks. (However, I've also recently heard arguments that the slower write times vs. read times might be an issue with this application. I'll dig a bit further into any downfalls with this use for SSD.)

He covers many other demos covering topics like active memory sharing, the Integrated Virtualization Manager, partition mobility, WPARs, nmon, etc. An index is available so you can browse around for the topics that interest you.

YouTube also has some relevant material. Search on "ssd vs. hdd laptop" and watch demonstrations where people put two identical laptops (save for the disks) side by side and boot them. Of course, these disks aren't the same as those in your Power server, but you can still see the performance gains--those laptops with SSD are much faster than those with traditional drives.

As the technology evolves and the prices come down, I expect more customers to examine SSD's benefits, in both their servers and their personal machines.

A Look at Recent IBM Announcements

IBM made some new announcements last week. Take the time to get familiar with this information.

From the overview:

"IBM announced new Power Systems blades and servers, virtualization software and POWER6 technology, all designed to help companies build, manage and maintain a 21st century dynamic infrastructure."

Highlights include:

• New 4-core JS23 and 8-core JS43 blades using 4.2 GHz POWER6+ processors.
• Virtual tape support for IBM i and BladeCenter implementations. One thing that customers wanted was a way to natively back up their IBM i images directly to tape drive -- this announcement will help address that issue. Customers will be able to use a LTO4 SAS-attached single tape drive model TS2240. Once they upgrade VIOS, firmware on the blade and PTFs for IBM i (in the May 22 timeframe), they'll be able to use the tape drive as if it were natively attached.
• Power 520 Express and 550 Express servers with POWER6+ processors. The 520 has 2- and 4-core 4.7 GHz models, and the 550 will have 2-, 4-, 6- or 8-core 5.0 GHz models.

According to the overview, "The POWER6+ processor has also been shipping in the 3.6 GHz model of the Power 560, the 4.4 and 5.0 GHz models of the 16-core Power 570 and the 4.2 GHz model of the 32-core Power 570 since they were announced last October."

In addition, IBM has announced 69 GB solid state drives (SSD) to be used with Power Systems. I hear that you should expect performance increases when you put your "hot" data onto SSD drives versus
traditional drives. Using SSD allows you to use fewer drives (removing mechanical seeks and waits frees you from worrying about spindle count). This reduces power consumption and takes up less physical space.

Another thing I hear is that, to get to the most out of the SSD drives, you want many random reads and a low percentage of sequential/predictable reads, and you want a higher percentage of reads than writes.

More from the overview:

"PowerVM, IBM's virtualization solution, is now enabled for the first time with Active Memory Sharing, an advanced technology that allows memory to automatically flow from one logical partition to another for increased utilization and flexibility of memory usage." IBM published a Redpaper titled "PowerVM Virtualization Active Memory Sharing" on this topic.

This gives you the option of telling your LPARs that they have more physical memory than they actually do. Basically, you'll be paging memory to a paging device. So if all your LPARs try to do this at the same time, you could face the same kind of contention and poor performance that you'd have on a single LPAR that is memory-constrained and constantly using paging space. Therefore, you wouldn't want to do this with LPARs that all need simultaneous access to memory. However, it is helpful if you have individual LPARs that need additional memory at different times. Think of this as another way to better utilize your server. With Micro-partitioning we can better utilize CPUs; with Active Memory sharing, we can better utilize our memory across our servers. Active Memory Sharing will work across AIX 6.1, Linux and IBM i 6.1 partitions that use VIOS and shared processors and have virtual I/O. Of course, they must be POWER6 processor-based systems with the Enterprise Edition of PowerVM.

Look over all of the announcement material, including the new facts and features documentation. Then contact your business partner or IBM to help you determine how this new technology can help your business.

Snapshots Make File Recovery a Snap

Watching this technical demo by Nigel Griffiths got me thinking. I like taking snapshots and flash copies on my external SAN devices. What if I want to take a snapshot of a file system on my AIX machine that isn't connected to a SAN?

Read more about these capabilities here.

http://publib.boulder.ibm.com/infocenter/systems/index.jsp?topic=/com.ibm.aix.baseadmn/doc/baseadmndita/jfs2_snapshots.htm

From the document:

"You can make a point-in-time image of a JFS2 file system that you can then use for backup purposes. The point-in-time image for a JFS2 file system is called a snapshot. The snapshot remains static and retains the same security permissions that the original file system (called the snappedFS) had when the snapshot was made. Also, you can create a JFS2 snapshot without unmounting the file system, or quiescing the file system. You can use a JFS2 snapshot to:

* Access the files or directories as they existed when the snapshot was taken.
* Backup to removable media.

"There are two types of JFS2 snapshots: internal and external. A JFS2 external snapshot is created in a separate logical volume from the file system. The external snapshot can be mounted separately from the file system at its own unique mount point."

So how do I do it? I'll duplicate what Nigel did in his video demo using my test machine that runs AIX 6.1.

First, I use the following command to create a file system in rootvg that's 500MB in size:

crfs -v jfs2 –g rootvg -a size=500M –m /snaptest -a isnapshot=yes

Then I mount the file system:

mount /snaptest

Then run two commands:

touch hello this is a test

echo "this is a test" > test.out

Finally, I run:

ls -la
drwxr-xr-x    3 root     system          256 Oct 23 14:42 .
drwxr-xr-x   22 root     system         4096 Oct 23 14:37 ..
-rw-r--r--    1 root     system            0 Oct 23 14:38 a
-rw-r--r--    1 root     system            0 Oct 23 14:38 hello
-rw-r--r--    1 root     system            0 Oct 23 14:38 is
drwxr-xr-x    2 root     system          256 Oct 23 14:37 lost+found
-rw-r--r--    1 root     system            0 Oct 23 14:38 test
-rw-r--r--    1 root     system           15 Oct 23 14:42 test.out
-rw-r--r--    1 root     system            0 Oct 23 14:38 this

To see the available menu options, go here:

smitty > system storage management > file systems > add change show delete filesystems > enhanced journaled filesystems

The menu options are:

List Snapshots for an Enhanced Journaled File System
Create Snapshot for an Enhanced Journaled File System
Mount Snapshot for an Enhanced Journaled File System
Remove Snapshot for an Enhanced Journaled File System
Unmount Snapshot for an Enhanced Journaled File System
Change Snapshot for an Enhanced Journaled File System
Rollback an Enhanced Journaled File System to a Snapshot

In our case we want to create a snapshot. This can also be done from the command line:

snapshot -o snapfrom=/snaptest –n testsnap

Running ls –la reveals no additions in the file system, but if you cd to .snapshot, you'll find the files you just created.

I cd into /snaptest/.snapshot/testsnap and run ls to find all of the files that were in my file system when I took the snapshot.

Creating a snapshot allows you to easily recover files should someone delete them, without having to resort to restoring from a TSM machine or some other backup mechanism.

Why Consider IBM Storage? Performance

Once the network and fibre cables are plugged in, many server administrators consider their work finished. Getting the LUNs needed to load their data or doing the magic that allows their server to talk to the world? That's what network and storage teams are for.

Of course, in some organizations, the server admin, network admin and storage admin are one and the same. But whether we're working in small shops or large enterprises, we must understand the impact storage has on overall system performance.

In the case of IBM storage, here's a handy link for your toolkit. The information here is self-explanatory, but I would direct your attention to that Web site's eLearning section. This section features the IBM Learner Portal, where you'll find material on topics like cabling machines, installing hardware and software, and configuring the DS3000, DS4000 and DS5000 storage families. There are also links to webcasts, classroom-based education and a mailing list.

So what does IBM have going for it in a competitive storage market? High-performing devices. Take a few moments to learn about the advantages of IBM storage here. Even if you have no intention of becoming a storage expert, knowing what to look for when you're disks aren't performing as expected will only make your job easier.