 |
 |
|
| Home > | Next > |
| | | | | |||||||||||||||||||||||||||||||||||
|
|
|
Fast Track to Solaris 10 Adoption: ZFS TechnologyGeneral InformationPlease click on a question below or download a pdf version.
Q: Who came up with the name ZFS, and what does the "Z" stand for? A: ZFS was originally the project name where the letter "Z" stood for Zettabyte, which indicated the virtually unlimited scalability of the file system. Now the "Z" no longer stands specifically for Zettabyte. Given the several advanced capabilities offered by this technology, the letter "Z" could stand for the "Zen File system" or "the last alphabet on file systems" or anything else. Q: When is the Solaris 10 OS slated for full release? A: The current schedule is to ship the Solaris 10 OS at the end of calendar year 2004. Q: Is ZFS based on the "Advanced File System," which is the base for SAM-FS and QFS? A: No, ZFS is not based on "Advanced File System." ZFS is Sun's next generation file storage solution, designed from the ground up to meet the modern needs of a general purpose, host-based file system. Q: What is the package name for ZFS? SUNWZFS? A: This is yet to be finalized, but internally we're using SUNWZFS for now. Q: Will ZFS replace the Solaris Volume Manager and Veritas Volume Manager? A: In the sense that ZFS does not require a separate volume manager, yes. If you're still using UFS or VxFS, ZFS will not replace their respective volume managers. Q: What are drawbacks in using ZFS, or situations that might not be suited for ZFS? A: ZFS works well in all situations. Q: Is ZFS the strategic direction for Solaris filesystems? A: Yes, for Solaris OS. A: Each block. In ZFS, we support multiple block sizes where each block can be from 512 bytes to 128K (maybe larger in the future). A: Yes. With access to all disks in a system (not hidden behind a RAID controller), we can provide functionality like self-healing data, priority/deadline scheduling, etc. Q: Any thoughts on porting ZFS to Linux, AIX, or HPUX? A: No plans of porting to AIX and HPUX. Porting to Linux is currently being investigated. Q: In a container environment, what is ZFS's role? Is there a ZFS for each container? A: Basically, the rules for ZFS with containers would be just as with any other file system; shared read-only is doable, multiple writers would be risky. You'd most likely have one storage pool on the system, with separate ZFS filesystems for each container. Q: How about the access control and security features for ZFS? A: ZFS will support NFSv4-style access control lists (ACLs), in addition to the standard POSIX user/owner/group modes. Q: Are there any plans to make ZFS a distributed file system in the future? A: This is currently being investigated. A: Yes. Q: How do you limit the storage available to a file system? A: ZFS offers virtually unlimited data scalability. ZFS is a 128-bit file system. It will provide about 16 billion-billion times more capacity than the currently available 32-bit and 64-bit file systems. In other words, ZFS is designed to support more storage, more file systems, more snapshots, more directory entries, and more files than can be created in the foreseeable future. Q: What's the thinking on using ZFS inside an Solaris Grid Container? ZFS on ZFS — NFS? A: From the container's perspective, ZFS is just another file system; it would be made available and appears just as UFS, QFS, and so on. Q: Do I need to use command mount in Solaris OS with ZFS? A: Simply by creating a ZFS filesystem, it will be available in the namespace (i.e., mounted). Of course, you will also be able to turn off this feature and manually use the mount(1m) command. Q: Does ZFS have multipathing load balancing i.e., powerpath/veritas multipathing/etc.? A: Yes. ZFS sits on top of multipathing capability MPxIO which is part of the Solaris 10 OS. Q: Is there a reason SAMFS and QFS are not available at install time? Will they be? A: Those products are additional cost items on the Sun price list, which is why they aren't available in the core install. Q: How are ZFS and QFS related? A: ZFS and QFS are separate products; although there's overlapping functionality, there's no shared code, for example. We see there being different places where each will play; this is briefly touched upon in today's sun.com cover story. Expect more guidance and positioning of different filesystem technologies as ZFS gets closer to release. Q: Will ZFS replace traditional volume management? A: In a word, yes. ZFS incorporates the functions typically handled by separate volume manager software, which makes for simplified system administration. Q: When will ZFS be released via Solaris Express? A: We are working hard to get ZFS into a build of Solaris Express prior to our general ship of the Solaris 10 OS, but that schedule is still not firmed up. Please join the Solaris Express program and stay tuned. Q: Does ZFS + Samba put Sun in a position to better compete with NAS vendors like NetApp? A: Yes, ZFS will drastically improve ease of management as well as performance. Among other things, this will improve Sun's position in the NAS space (both with NFS and Samba). A: ZFS is not in Software Express at the current time. Q: Is ZFS available separately? Can I convert from HFS to ZFS on the Solaris 9 OS, for example? A: There are no current plans to offer ZFS on previous versions of the Solaris OS. You will need to install the Solaris 10 OS. Q: Is ZFS in the base Solaris 10 OS, or will it be additional cost? A: There are no current plans to charge for ZFS. Q: Why go to ZFS? What does it offer compared to UFS and VFS? A: ZFS offers a number of enhancements compared to traditional filesystems. You can find an overview on the sun.com feature story. A: ZFS will be in the Solaris 10 OS when we ship the product. The current projection is the end of 2004.
| | |||||||||||||||||||||||||||||||||||
