Enterprise Unix Roundup: Of Blogs and Bottom Lines Page 2

By Michael Hall (Send Email)
Posted May 19, 2005


Main     In Other News     Recent Updates     Tips of the Trade

In Other News

» Open source pundit Doc Searls has floated the idea that it might not be outlandish for Microsoft to be sizing up Red Hat for acquisition. Bong-hit conspiracy theories abound along with token snark from the "Red Hat is no better than Microsoft anyhow" contingent. Our take? If it were to happen, we'd be quick to glue some glitter on a shark and mail it to Redmond as the first annual Enterprise Unix Roundup "Stunt Acquisition of the Year" award. Then we'd sit back and watch the beard-and-suspenders set exodus to Novell and SUSE.

» Montavista released Carrier Grade Linux Edition 4.0. The company says it's the only carrier-grade Linux distribution that meets the OSDL's Carrier Grade Linux 2.0 requirements. As well it should, considering Montavista is an OSDL participant and contributed several members of the group that wrote the specification.

» SCO announced a whitebox partnership with DTR Business Systems and Terian. The two will fill orders for SCO OpenServer, UnixWare, SCOoffice Server, Microlite Backup, and VSI/Fax Server systems, placed by resellers on the SCO Web site. SCO also filed a prospectus with the SEC yesterday, duly noted and dissected by Groklaw. Upshot: The company is plagued by short-sellers and a dead-in-the-water licensing program. Its remaining eggs are all rattling around in a litigation basket.

» Linux clustering vendor Scali on Monday released the latest version of its Manage and MPI Connect Suites. New features in version 4.4 center around storage management functionality and allow multiple operating systems to coexist in a managed cluster. The new version integrates storage management with overall cluster management and includes additional cluster availability and reliability features.

» Fujitsu gave five of its 64-bit SPARC-based PrimePower servers a speedbump this week. The 650, 850, 900, 1500, and 2500 now contain 2-GHz, SPARC64 processors that bump clock speed as much as 14 percent and Level Two cache by 33 percent. Enterprises looking to upgrade on the cheap can swap out the chip set, Richard McCormack, senior VP of Fujitsu Computer Services told ServerWatch. He also noted that enterprises can mix and match processors within a CPU. This release, McCormack, said, "keeps us [Fujitsu] a very strong choice, right there on the short list with IBM and HP."

Recent Updates

  • OpenBSD version 3.7 was announced today. The project has provided a handy change log that details numerous bugfixes, improvements, and additions to its supported hardware. OpenBSD is distinguished for its rigorous security testing: "We remain proud of OpenBSD's record of eight years with only a single remote hole in the default install," said project founder Theo de Raadt.

  • Apple released OS X 10.4.1. The point release, out mere weeks after the general release of OS X 10.4 (Tiger) provides much in the way of client-side enhancements and fixes, and it addresses the minor furor regarding widgets of mass destruction, a potential exploit Tiger's new Dashboard feature introduced through its modification of Safari's behavior.

  • List server Lyris was updated to version 8.5b. The update includes minor tweaks.

  • Mail server CommuniGate Pro was updated to version 4.3.1, providing bugfixes for version 4.3, which arrived two weeks ago.

  • Mail server Postfix was updated to version 2.2.3. The changelog lists two bugfixes.

  • Kerio Mail Server was updated to version 6.0.10, providing 10 bugfixes.

  • FTP server CrushFTP was updated to version 3.9.1.

  • The Squid Web caching proxy was found to contain an error in the way it handles DNS validation, which leaves it open to potential DNS spoofing attacks. Patches have been released by Ubuntu, Red Hat, TurboLinux, and Trustix.

Tips of the Trade

Logical volume management is one key to sys admin happiness, particularly for sys admins faced with rapidly growing and changing data storage needs. The Enterprise Volume Management System provides a single unified system for managing all locally-connected hard disks on a Linux or Unix system. With EVMS, you can easily scale up data storage needs as you go and not be hobbled by old partitioning schemes or physical drive capacities. Just add new drives and create virtual volumes that span disks, painlessly adding and changing storage space as needed. EVMS can easily be added to systems; it's not something that must be built in to the network from the very beginning.

EVMS supports clustering and recognizes most filesystem types, including Windows, Macintosh, S/390, RAID arrays, and most partition types. It includes some very useful features, such as:

  • Snapshotting, for creating images of volumes. This is useful for backing up data because the snapshot is backed up so the original volume need not be taken offline. Database admins should especially appreciate this feature. Snapshots can also be used to restore a system to an earlier state.
  • Drive linking, for combining EVMS objects to create a larger contiguous volume. Drive links can be shrunk or expanded as needed.
  • Filesystem management utilities, such as creating and checking filesystems on volumes.
  • Bad block relocation.

All of this adds up to high availability at low cost: EVMS is cost free, so the only items to purchase are hardware and time. Learn more about EVMS at the EVMS home page.

Carla Schroder writes the Tips of the Trade section of Enterprise Unix Roundup. She also appears on Enterprise Networking Planet and Linux Planet, covering Linux from the desktop to the server room, and is the author of the Linux Cookbook.

>> To Main

Page 2 of 2


Comment and Contribute

Your name/nickname

Your email

(Maximum characters: 1200). You have characters left.