Slashdot Mirror
Is Storage Capacity Outstriping Backup Capability?
Kzip asks: "On my modest home LAN we have four computers with around 300Gb of storage. A lot of this is used, but not a lot of it is backed up (certainly not on a regular basis). When I started looking for a backup solution I found that most of the affordable tape backup was way to small (DAT 12/24 is just too small now a days) or too slow (Onstream do 50Gb but on IDE it's only ~1MB/s ... so 6 tapes over 80+ hours!) or just too expensive (HP Ultrium is great, but at £3000 for a drive and £120 per tape it's a little pricey). So I'd like to ask the /. community: Does anyone know of a fast and affordable backup system for home/small office use." After a quick scan of Pricewatch and other sites, it seems that backup solutions >99G are expensive (all the ones I could find were more than $1000US). How long will it be before these and
terabyte-backup solutions become affordable for SOHO ? use?
At my university, our college of engineering bought a StorageTek Powderhorn for interdepartmental backup. The model we have currently has 100 TB of storage capacity and can be expanded to 300 TB. Its host is a massive Sun server connected to the core network switch via two gigE links and and one ATM link. At the server level in various departments and groups we are mostly doing RAID as disks have become so darn cheap. A simple script dumps data onto the Powderhorn across the street once a week in the event of a major malfunction (RAID recoveries don't always go smoothly), theft, or fire.
I got to thinking once again about how fragile my current backup scheme is.
I have three machines -- one laptop for MS Office, games, and browsing the rare website I can't use with Opera on Linux, one ancient Linux box doing NAT and email, and one modern Linux workstation that I use for my daily work. My backup sets from these machines total about 15 gigabytes compressed. About five gigabytes are irreplaceable data, and replacing the rest would be several days' work.
I take backups every night. The two linux boxes cpio, gzip, and cp new and changed files off their relevant filesystems to a separate backup drive on the workstation, while the built-in backup applet on the Win2k box takes a full backup onto the same drive via SMB. An additional cron job on the workstation renames the windows backup file each night so the new one doesn't.
This scheme protects me from any single drive failure, as well as accidental deletion of any file or directory except the root directory on the Linux workstation. Which means that if, for example, if I were to install a poorly constructed RPM that did rm -rf $DIRECTORY/ in a script while $DIRECTORY was unset, I would lose all my data. While I do have the most critical stuff on CD's, those are usually weeks out of date. And even those would be susceptible to a housefire or a burglary (well, I don't know if burglars typically bother to steal stacks of CD-R's -- anyone with experience in this regard?).
In any case, offsite backups would be the way to go. CDR's are a pain to automate when your backup set is large, and a pain to drag off site, and they can get quite expensive in the long run, so I've looked into various online backup services. Unfortunately, at the volumes I'm looking at, the commercial services will typically charge you several times the one-time cost of the hard drive space each month, which seems somewhat excessive.
In fact, if I could just find two or three random guys with DSL or cable, each with a server that's always on and has 20GB to spare, I'd be quite happy to give them each 20 gigs on my box in return. A 15-minute search on Google didn't turn up any backup exchange rings on the web -- is someone doing this kind of thing, either privately with friends or through a more open group? What kind of software do you use? While I would be perfectly happy with cpio | gzip | gpg | ssh cat to stash my own stuff, I would be hesitant to give random strangers full shell accounts on my box. And I would prefer not to let them turn my workstation into a warez server, either, although I suppose IP address restrictions and monitoring would pretty much take care of that. Something that runs on windows would also increase the user base nicely.
Has anyone been thinking carefully about a peer-to-peer online backup system, or should I?
There is a backup solution by a company in Seattle called InterVault. It allows you to backup your data to their servers via the Internet. They use encryption to protect your data, so don't lose your password. Their is no backdoor. Check them out at www.intervault.net.
-> Capt Cosmic <-
So you think you're pretty smart by buying two big HDs and doing a nightly copy of one to the other? Just think what happens when the source disk fails in the middle of the nightly backup. You have a failed source disk. And a half-baked backup disk. With a possibly unrecoverable file system. You just lost all your data.
I think if you want to do backups to HD you need three of them!
All the Unix backup tools can backup to disk as easily as to tape. Carriers to make ATA/100 disks removable cost about $10 each. ATA/100 disks are cheap per megabyte.
There are techniques to make the disks hot swappable, or use a dedicated backup machine that can be easily powered down to swap disks.
Most importantly! It's a restore system, not a backup system.
Nobody cares how great your backups are, if you can't do a restore when you need it.