Archive | April, 2007

DLSLUG notes, 5-Apr-2007: Todd Underwood discusses ZFS

Fifteen attendees made it to the April meeting of the Dartmouth Lake
Sunapee Linux User Group
, held as usual in Carson Hall at the Dartmouth College campus on the first Thursday of the month.

Todd Underwood, Vice President, Operations and Professional Services at Renesys presented “ZFS: The Last Word in File Systems.” Renesys is in the business of collecting, analyzing and archiving data about what’s happening on the internet and, not surprisingly, that’s a lot of data and growing geometrically. One of Todd’s projects is to provide fast and reliable storage for the hundreds of gigabytes per day acquired and the tens of terabytes of data stored. He presented a survey of what’s out there, what his needs were, and how he reluctantly narrowed the search down to SUN Solaris and ZFS. While he had nothing but praise for ZFS, he expressed some reservations about SUN. Strong reservations.

Todd dug into the ZFS architecture. ZFS is truly amazing: disk contents are always coherent, with writes all checksummed, writes as atomic transactions, with “fancy FS internals” like IO scheduling, dynamic block sizes and prefetch queues, huge limits (128-bit data). ZFS flattens the Linux file system model of multiple layers into a single monolith, and eliminates entire classes of problems introduced by the multiple layer architecture. There is no fsck. Devices are accumulated into pools. File systems are assigned to pools. Storage addition and maintenance is fairly trivial. ZFS performance is remarkable, at disk speed (with compression, sometimes in excess of disk speed).

Todd presented the rough outlines of his storage system: $17k worth of hardware from SUN and Dell yields 7.5 terabytes of storage with nearly a gigabyte per second throughput. Todd’s disk storage challenge is solved, as the company’s demand can’t match the throughput capacity of his system, for now, and the system can be expanded with additional external storage.

As for the porting of ZFS onto other architectures, Todd expressed the opinion that running on Solaris or OpenSolaris is likely the best current solutions. Porting onto BSD is underway but not yet ready. He had heard that ZFS was ported to OS X but could not confirm (Googling ZFS OSX yields interesting results).

Todd promises to send along slides from his presentation; I’ll try to post links to them to the website.

After the main presentation, there was a good session of questions and answers. I asked a question on replication of Postgres and Todd recommended the Sequoia product.

Thanks to Todd for his thorough presentation, to Bill McGonigle for organizing the meeting, to Bill Sconce for sharing his notes with me for this post, and to all for attending and participating. Next meeting, still tentative, will be on writing a FireFox extension by Roger Trussell.

Notes from CentraLUG, 2-April-2007: Bill Stearns on LVM

A great meeting last night. The Central New Hampshire Linux User Group met at the usual place and time: The New Hampshire Technical Institute‘s Library, Room 146, 7PM on the first Monday.

I did my usual rounds of announcements, Shawn O’Shea pointed out that besides for the discuss and announce mailing lists, you can also subscribe to the lists via RSS by using one of GNHLUG’s archival sites. Add one of these to your favorite RSS readers to see the GNHLUG announcements: mail-archive.com or gmane.org

Everyone got to introduce themselves and speak a little bit about what they’re up to. I passed around a couple lists of topics and speakers from the wiki to find out what the attendees want to see for future sessions.

Bill Stearns presented “LVM: Logical Volume Management.” He explained about the basic need to expand or re-allocate disk resources without making hard partition changes, in some cases without even shutting down. We started right in on an exercise: using the loopback device and some spare space in /var/tmp, we created three loopback block devices. We assigned them as PVs (Physical Volumes) and allocated two to a Volume Group. Then, space could be allocated out of the volume group to provide the space needed. Additional PVs could be added to the VG, additional space from the VG could be allocated to a mount. We had a good discussion about the choice of filesystems and the different processing required for ext3, reiser, xfs, jfs file systems – most of which can be resized when on-line. (Bill recommended the MythTV HOWTO for a good discussion of which file system to use.

We had a good side conversation at this point about the mount tables and the significance of several flags Bill had. An intesnse discussion of ‘noatime’ – useful for hierarchical file management, but generally of little use, and a lot of disk activity, power consumption, speed decrease, and for FlashROM devicers, perhaps lifetime shortening can be avoided by adding noatime to the mount tab. Bill also had war stories about the security implications of nodev and nosuid both of which are a good idea for insertable media unless you have a specific reason for needing them.

We finished up the LVM exercise by adding the third PV to the VG and then resizing the ext3 partition to include the space. Bill took questions from the audience: one lady had just installed LVM on the Linux partition on her mainframe (!) that weekend, and wanted to know more about LVM striping. Other questions on reliability, use with RAID. Bill had some pointers for adding additional storage: use of USB2 (not USB1!) external drives (Bill hasn’t been happy with Firewire storage on Linux) or using an external storage solution (he mentioned CoRAID which uses rackable ethernet-to-ATA raw drives; Bill had one sample to play with) and handled some questions further afield, like the file defragmenter Bill has on his site.

Yet another great LUG meeting. Thanks to Bill Stearns for the great presentation (and 3-ring-bound handouts) and providing the raffle door prizes. Thanks to Bill Sconce for providing the projector and doing the note-taking during the meeting. Thanks to the New Hampshire Technical Institute for providing the facilities.

Look forward to a great presentation on OpenWRT by Ben Scott at the May 7th meeting (where you can expect him to be heckled) and another great presentation on Drupal by Seth Cohn on June 4th. Hope to see you there.

Life After VFP

Robert Jennings posts Yet “Another Life After VFP Thread.” For those not following VFP closely, MS recently announced a confirmation of earlier news that there were no plans for a VFP version 10, and that the VFP scripts in the project known as Sedna would be released under some sort of public license. Poor communications lead to media and Slashdot reports that VFP was to be Open Sourced, sadly not the case.

Robert does a good job of outlining the huge cost in moving a vertical-niche application into another development environment, language and runtime. Most sophisticated specialty applications have person-years of investment built into them, knowledge not easily extracted, transferred or translated to any new environment. Regardless of whether that new environment is Dot Net, Dabo, LAMP, Python or Visual Fred, there will be a huge cost and risk with any enterprise making this switch.

Unlike the Open Source world, when a vendor choses to discontinue a product, developers have little choice but to move along. While many folks point out the upside that the product will likely run for years to come, and a lack of Microsoft official support doesn’t instantly obsolete a product (DOS apps can still be found, after all), there is an immediate slowdown in the custom software market, and a longer-term turning away from the product by customers. Large-scale vertical products have to be operating with 5- and 10-year plans for reinvestment and changes in direction, to ensure they can fund “The Next Big Thing” while continuing to deliver good value to their customers today and tomorrow.

This is not a death knell for the product. The writing has been on the wall for years. But developers with large applications have to be looking around for a new platform.

FoxPro developers always viewed themselves with a bit of “Battlestar Galactica” mythology: a rag-tag crew of self-taught developers from the PC Revolution, they survived the dBASE wars and the implosion of Ashton-Tate. Working under a cruel master who never promoted their product, they persevered. MS’ internal team developing VFP did amazing things on a shoestring budget, introducing a fairly smooth transition from procedural to object-oriented, from developer-guided to event-driven interfaces, from characters to pixels, from local ISAM to RDBMS. The VFP IDE was a remarkable environment in which to develop rich-client, component-based, web-driven or even server-based applications. I will miss it, and look forward to becoming as skilled at my next platform.

Powered by WordPress. Designed by Woo Themes

This work by Ted Roche is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 United States.