Slashdot Mirror

Ceph is a filesystem that distributes over known nodes and auto-rebalances when new members are added or removed as well as overbalancing "hot" reads. It may be interesting to see how many nodes one can add.

I'm probably pushing 1TB of photos on DreamHost. They're maybe 98% uptime but for something like I use it for it really doesn't matter. For what I pay it's great.

Plus one of their employees wrote Ceph. (A FOSS distributed file system).

Could you also add Ceph to your wishlist ?

http://ceph.newdream.net/

thank you. :-)

Plus they wrote Ceph, (distributed/scalable file system, which merged into 2.6.34.)

Where did you find that information? The Ceph page suggests that it originated out of a University research group:

This project is based on a substantial body of research conducted by the Storage Systems Research Center at the University of California, Santa Cruz over the past few years that has resulted in a number of publications.

Furthermore, it makes no mention of Dreamhost whatsoever.

I have Chrome, Thunderbird, my MP3 player and DropBox on TrueCrypt partitions.

Computer is PowerCycled and it's "gone". Since speed isn't a huge factor I went paranoid and went with AES-Twofish-Serpent. Good luck recovering my stuff.

I use DreamHost for my mail/webserver. They're not 5-9s but they're cheap and still seem like they are a "small company". Plus they wrote Ceph, (distributed/scalable file system, which merged into 2.6.34.)

I'm sure you could write cron script or something to run on the shell to do what you're talking.

I'm not gonna go through and do a feature check, but there is Ceph. It's still pretty early in the development, but looks pretty promising. It uses btrfs as the underlying filesystem.

Someone put together 1.2PB of Glustre (with dual replication) at my company and it's been problem free so far...

The headline in the Ceph wiki: Ceph is under heavy development, and is not yet suitable for any uses other than benchmarking and review.

ceph, XIV, and other distributed storage controller models are available today, and avoid controller bottlenecks.

The first thing you need to know about RAID5 is that it's pretty unreliable; if you lose one device (and subsequently replace it) then the array has to read every sector from every other device in order to rebuild the data. Any unrecoverable sector error on any device will result in a corrupt sector in your rebuilt array.

RAID1 duplicates devices, although your storage requirement is now 2x the quantity of data being stored (as opposed to say 1.25x), the chance of error on rebuild is a lot smaller.

However, all inexpensive RAID solutions suffer from the problem that your devices are on a single server - they're a single point of failure, and if, for example, your server's power supply fails and fries the parts in the case, all copies of your data may be destroyed.

To mitigate that problem you could try a distributed filesystem. Your files would actually be distributed among multiple servers and the filesystem would ensure replication. MogileFS is one such, although it does not provide a POSIX filesystem view it is nevertheless pretty easy to use. There are various distributed filesystem projects around, including Ceph, Kosmos, and Venti.

Although these projects are at varying stages of completeness and you may need to be a bit brave to trust them with your important data, the promise of distributed filesystems is high availability and extensibility.

genious?

I find it too bad to still see so many on the tracks of faulty and damaging prescriptive grammar proselytizing. It's like pushing a failed political ideology long after such an ideology has proven itself untenable (Communists, please raise your hands, yes you know who you are). Anecdotal experience has shown that most prescriptive grammarians don't even know the source of their beloved rules. Allow me to enlighten. Most rules of proper English usage and spelling were developed during the 1600s to 1700s. The need to do this stemmed partially from the understandable need for a consolidating influence on the many various dialects (yes I said dialects) of the English language. It is instructive to look at the roughly comparable example of the Chinese Empire. When the Chinese faced the problem of spreading official law and philosophy throughout polyglot masses, they realized they could never force every citizen to abandon their perfectly functional native tongue for the "Emperor's own". Instead they enforced what they could, writing. They codified and standardized their ideograms and learned people used the official imperial version of the script. Even if the local populace used words differently in their dialect from the written form, it was considered proper and correct to write one way and speak another. Because of China's vast influence throughout Asia, the remnants of this system of education and learning are still with us. The Japanese still use Kanji, Koreans have Hanja and until the French introduced a heavily modified Latin based system, the Vietnamese used Hantu. Now, this analogy is a bit loose, but more or less accurate. The solution was to thus model English grammar on the language of the West's version of China, the language of the Romans -- Latin! Yes, that is correct, Latin. A few language busybodies (such as Bishop Robert Lowth ) had the bright idea that we should take principles from Latin grammar and apply them to a largely Germanic language. This is why we have rules fraught with so many errors and awkward "sounding" results. Most people of decent linguistic education laugh at these rules when they are brought up under such flags as "correct English" or "proper usage". This aged and archaic Enlightenment linguistic philosophy has resulted in such preposterous and often humorous rules such as: "don't split an infinitive", "don't end a sentence with a preposition", etc. Creating rules for one language based on rules for another is akin to throwing a woman in a river to see if she is a witch. Due to the same group of ignorant prescriptivists (http://www.techwr-l.com/techwhirl/magazine/writin g/grammar.html), spelling in turn became a complete disaster (http://en.wikipedia.org/wiki/English_spelling). I personally attribute much of the awkwardness of modern English spelling to the stubborn refusal of prescriptivists to introduce a letter for the schwa sound (which FYI is statistically the single most common vowel sound in the English language yet we do not have a single letter for it). Every vowel is at some point used as a placeholder for the schwa sound in unstressed syllables. Ask yourself sometime what the homophonic overlap is between any two vowels (I mean, common, the letter "o" in "women" has the same sound as "i" in "is" for goodness sake!). That's just for the vowels! English consonants are an absolute disaster. There are many reasons for this, most often attributed to the incredible number of loan words in the English lexicon and the representation the source languages uses for those sounds. Even more distressing are the wide variety of representations for common homophones. Take for example the common construct "ough". Let's play a game and see how many different sounds "ough" can represent: "cough", "rough", "through"

Sampling is a big deal, because you are using a chunk of the master recording, which is the end-product of many, many dollars of recording, mixing, mastering, promotion, artist development, etc. I worked for a while in New York with the guy who produced Rump Shaker and he had an attorney that dealt with all of his sample clearances. Moreover, there is no standard licensing charge for samples. The owner of the master recordings can charge you whatever they want, or not let you use it at all, because it's their product, and using their product to make your own product is not "fair use", it's profiteering.

On the other hand, re-recording a piece of a song is a lot cheaper and is subject to compulsory mechanical licensing, meaning that you pay 8 cents per re-recorded song for every CD you sell. Nobody can stop you from doing that. Dre did that with many P-Funk tunes on The Chronic.

It's one thing to re-interpret someone's idea. It's another thing to appropriate their implementation - the creator has every right to control that as they see fit.

I wrote and presented a paper a few years back about how digital communications could allow more communal input similar to the ancient Greek Agora. Seems like this is compelling evidence that it works (to a degree--I certainly don't expect Pinkerton to respond the way /.ers want them to; but, nevertheless, they're listening). One wonders, though, if this isn't some sort of dubious sympathy PR stunt on the part of Pinkerton. We'll have to wait and see, I guess.

---------------------------------------------
Yes! Oh yes! My soul is snoring! - Tom Servo