As my final tech demo of the year, I demonstrated ZFS (zettabyte file system: name now deprecated), the "world's most advanced file system" by Sun Microsystems. Out of many possible topics that I could have given a tech demo on as a Sun Campus Ambassador, I chose ZFS because for years (even before I had cared at all about file systems) I had been hearing about how "amazing" ZFS is.
And after having had the experience reformatting my hard drive, using external hard drives, reinstalling multiple operating systems, needing to install a new operating system in parallel with my current operating system, finding that operating systems like using different types of file systems, and that certain file systems have very annoying limitations, ZFS started to sound very interesting. I had to at least find out what it is.
In the process, I discovered Solaris; specifically, Solaris Express Developer's Edition (SXDE for short). ZFS was created to be the be-all-end-all filesystem, but also originally as a proprietary software offered exclusively by Sun Microsystems as part of the Solaris operating system. Now, Sun Microsystems has completely open sourced the ZFS code, along with a majority of the Solaris operating system core code.
I fell in love with both Solaris and ZFS. It took me longer to download Solaris than it did to get it "installed", booted up, and running on VMWare on my OS X machine; Sun offers a disk image for your MacBook Pro that is ready to go with the latest version of VMWare, all set up, installed, and ready to go. (Pictured above.)
Believe it or not, to get ZFS up and running, was even less effort. With a few simple commands, you can build a RAID-Z from a "pool" of "volumes". The best way to think about ZFS is like thinking about RAM. When you install a piece of ram into your computer, you now have an addition to a "pool" of memory which applications can store information temporarily. It makes your computer faster and better. You don't have to worry about "partitioning" the RAM, or how much memory you can store on an indiviual RAM stick. It goes into a "pool" of memory. If you had 2 GB, and you install another 2 GB stick of memory, you now have 4 GB of memory. Done.
ZFS works in a similar way. Storage space is no longer limited by the size of a single hard drive. You can "pool" multiple hard drives into one file system. You can create a RAID-Z from several volumes, and then add the RAID-Z to the pool. Long story short, if you have four 30GB hard drives in your sytem, you can "pool" them into a single file system that is 120 GB in size. If you have a 1GB USB stick, you can even pool that into the file system. Old barries simply don't exist anymore. Storage works the way it should.
Other noteworthy features:
- "All or nothing write" a.k.a. "transactional file system": Your file is written, and its data is in integrity, or not. There is no in between.
- "Predictive self-healing": Long story short, your file system fixes itself if it encounters errors. There's no need to run "scan disk"; if there's a problem, it's taken care of on the fly.
- Easy quotas: Want to set a limit (or a minimum) amount of space a user has control of? One command will do it.
- Safe delete: Files can be REALLY deleted -- as in, normal deletion of files means that the file entry is removed from a table of contents. This is similar to deleting an entry in the index of a book, so that you will never be able to find it again. But if someone actually took the time to read the book without the index, they would still be able to find the information -- this is the same for hard drives. Even though you think you deleted something, it is likely still there, and if someone stole your hard drive and wanted to recover the information, they probably could. ZFS allows you to actually delete the information where it was stored, as well.
- Incremental Backups or "Snap Shots": Back up your file system in one command. Back it up again tomorrow in one command, but only saving the differences between yesterday's and today's file system. This means you save a huge amount of space, only storing the differences between backups.
- 128-bit File System: Long story short, this basically means that you can never run out of space. There is enough "address space" to catalog every star in the universe and then some, with a 128-bit file system. Yes, the universe.
- Compression: Activate compression in one command. Hard drives are still limited by the speed in which the platter turns, while CPU speed is increasing exponentially. Now, we can use the extra CPU power to reduce the amount of data that is written to a hard drive, effectively greatly increasing our data throughput.
This was my last "official" Campus Ambassador tech demo of the year, though this week, I have the Semester Project: Final Review event--a program I organized that is sponsored by Sun Microsystems. It's been an amazing semester with Sun Microsystems, and I'm sad to see it coming to an end... I'm sure I will wrap up this blog with a few final thoughts on my experiences as a Campus Ambassador. Stay tuned.
hi im swapnil i want know about sun solaris is there any test requried i want know about administratative level beceause i give that exame so reply us
thank you..
Posted by swapnil vilas jadhav on June 15, 2008 at 01:23 AM PDT #
http://opensolaris.mirror.garr.it/mirrors/opensolaris/latest.iso
Posted by Deepak singh on August 07, 2009 at 05:23 AM PDT #