What is Digg?
52 Comments
- inactive, on 10/10/2007, -2/+10S3 is a ***** godsend.
- andrewcsayer, on 10/10/2007, -7/+15Don't read anything about bullets or them being fired at a server.
Buried. - lencioni, on 06/19/2008, -3/+8This is a good idea. The only trick is you can't be too lazy that you don't actually ever get around to setting up the automatic process ;)
- tpink, on 10/10/2007, -0/+5He's talking about backing up a personal website and database with S3. He never said it was meant for backing up a Fortune 500 company's datacenter.
- Firehed, on 10/10/2007, -0/+4Unless you manage to ***** something up to an unusually fantastic degree, the worst that happens is that your backups are no longer automated.
Though in honesty, the type of people that would set up something like this are, more likely than not, the type to know how to avoid getting a virus in the first place. - rudy23, on 10/10/2007, -1/+4nothing. they are on the server. the backup program only ulpoads updated data so it should still be there
- MalDON, on 10/10/2007, -0/+3I always thought s3 was expensive, but I guess not. After seeing that backing up my entire site would only cost me about 63 cents, wow.
- s0m31john, on 10/10/2007, -4/+7I can haz bucket?
- dhbanes, on 10/10/2007, -0/+2Are you backing up 160GB of unique data every night, or does your backup consist mostly of the same unedited files from day to day? You might consider switching to a system that utilizes delta encoding, like rsync. I'm sure this isn't a viable option with your tape backups, but you could probably employ a more convenient online system that takes far less time, is far more secure, and doesn't cost much more than replacing tapes.
- monosyth, on 10/10/2007, -1/+3www.jungledisk.com uses Amazon's S3... i'm very pleased. very cheap reliable backup everynight at 11pm (or whenever you set it up for)
- ferrix, on 10/10/2007, -0/+2Dugg, but I may be biased because I wrote s3sync =P
- Touchdown, on 10/10/2007, -0/+2I've been using the S3Fox plugin for firefox mentioned in the article and it's great with S3
- univerio, on 10/10/2007, -0/+2Anyone got a gun? Anyone bored enough to try it?
- eric0213, on 10/10/2007, -0/+2Setting up the automatic process is easy. Checking that it's running periodically and making sure you can recover from your backups is another thing all together.
- KIERANMULLEN, on 10/10/2007, -1/+2mozy.com offers cheap backups too. I just signed up for a free 2 gb account. I saw the link in the story.
- mateo60, on 10/10/2007, -0/+1How odd. When I needed something restored, it took about 2 minutes for them to send me the link.
- JasonHilton, on 10/10/2007, -0/+1oh nice, somehow I missed that. Thanks
- thewump, on 10/10/2007, -0/+1In the config file there are server login, dbname and password - so maybe it's handled. I wondered that too. Worst case I guess you could put a cron on the server to create a daily sqldump which would then be backed up with other files?
- inactive, on 10/10/2007, -0/+1This is a fairly cheap route, if you don’t use too much bandwidth. If you get up to around 2TB/mo or so, the normal bandwidth quota prices for server rentals, it ends up costing $360/mo whereas I can rent a 2TB/mo server for $90/mo.
- motheroats, on 10/10/2007, -2/+3i has a bucket
- dezmd, on 10/10/2007, -0/+1Ok, given, its geared towards more 'personal' data, but I'm fairly certain this is akin to a press release, its what 'new media' has become. I've seen these services being advertised, extermely cheap even, for YEARS now, this is absolutely nothing new.
- mduke, on 10/10/2007, -0/+1I subscribed to the mozy unlimited service. Its only $5 a month for unlimited storage and seems to work great so far. Its real fast but only works with windows and Mac so far. I recommend it.
- MariusVW, on 10/10/2007, -0/+1Great solution.
Currently I have 3 options:
1) Amazon S3 / s3sync
- Great API... you can write all sorta of easy custom Applications
- No pre-existing Amazon way to browse files... One needs to resort to a great Firefox plugin though (Until Fuse s3fs comes)
- Great bandwidth
- You pay for everything... scalable cost (but cheap)
- You have to fiddle with s3sync and ruby and cron to get it working.
2) Mozy (mozy.com)
- Great *date versioned* incremental backup/restore solution.
- Not very flexible at all. (Can't use it as a generic store)
- Still a bit buggy, but shows great potential
- Nice web-based restore
- Nice "unlimited" storage for $5/m (don't trust this, but it seems to be true)
3) rsync/ssh (like s3sync but)
- For this you need your own backup host/internet connection, so cost can vary
- Ultimate freedom and flexibilty. (but not the reliability of S3)
I've been trying out this S3 thing for a bit now, and it looks like a great solution actually... may work out cheaper than my current Mozy subscription. (My rsync is also still chuggin away, but I don't trust my backup host either...)
So for now, S3 yay! - Zer0Fade, on 10/10/2007, -1/+2Mozy ***** sucks! I used it for a month and the one time I needed to restore something it took 3 days to process. ***** ridiculous!
- Everglow, on 10/10/2007, -0/+1Don't digg zer0 down; it's a legitimate experience. Was there any response from their tech support for the length of your data delay?
- m2paper, on 10/10/2007, -0/+1I can haz bucket?
- opensourcemaven, on 10/10/2007, -0/+1it does that too, look in the script where it has mysqldump -u...
- ericnmu, on 10/10/2007, -1/+2you forgot tape backup too
- opensourcemaven, on 10/10/2007, -0/+1depends what you're backing up. he was talking about his 2yr old blog, where all the files on his server amount to 60mb compressed. he mentioned he's on a dedicated media temple server, after googling i found out his box is a dual-core xeon.
- Firehed, on 10/10/2007, -0/+1Although you can bet that everything on Amazon S3 is pretty damn redundant. That's like assuming that only one copy of a message exists in Gmail. When you've got tens or hundreds of thousands of servers, as I'm confident Amazon does and know that Google does, things get replicated dozens of times. Think of it as a massive RAID5 array of RAID5 arrays of whole systems, each themselves probably using a RAID5 array (though it would probably be a mirrored RAID5... 15? 51?) As I don't work for either company I can't comment on the actual logistics, but it's a pretty safe assumption that not much short of a nuclear holocaust would completely eliminate data stored at Amazon.
- makjaksoftware1, on 10/17/2007, -0/+0this is very interesting information given. If you want to know about how can you buy Latest Software and OEM CDs in cheap price, then you can visit http://cheap-software-oem.blogspot.com/
- OBKenobi, on 10/10/2007, -1/+1You cannot has bucket.
I also fail to see why you'd want to back something up to an online server. Unless you're concerned about natural disasters. Yes, I suppose it might be helpful for Iraqis, when faced with having their datacenter blown up on a daily basis.
Bulletproof server... ha! - opensourcemaven, on 10/10/2007, -1/+1no, i just forgot about the reply button.
- lionslair, on 10/10/2007, -1/+1delete wrong topic sorry
- sancho, on 10/10/2007, -0/+0A nice midground would be to store a hash of your file in metadata and compare metadata to a live checksum on the box before backing it up. If the checksums match, there's no need to back up that file.
Of course, if you have lots of files with small deltas, you'll still be backing up the entire file to S3. Bottom line? Be sure to know what sort of backup scheme is best for you. - mishaco, on 10/10/2007, -1/+1bullet proof server backups . a dream if there ever was one .
- rudy23, on 10/10/2007, -2/+1even the most basic backup software gives you the option to do incremental, diffrerential and # of full backups to store option.
- EnterDaMatrix, on 10/10/2007, -2/+1S3 should be journaled too so your data from a few days back should be safe. If it doesn't keep separate instances of your data than in my opinion it isn't a service worth using.
Also Digg now has this great reply button. - dimitry, on 10/10/2007, -4/+3Talk about detailed and helpful. Awesome stuff Paul as always.
- tehmoth, on 10/10/2007, -1/+0this would be more useful if it said something like 'bulletproof hosted shell account backups'
- zombieball, on 10/10/2007, -2/+1I think he wanted to be cool and have his post read by all rather than burried under the previous post (just like mine is right now) :P
- vonskippy, on 10/10/2007, -2/+1What kind of lame ass home based server are you talking about?
I backup about 160G per night - exactly how is that going to happen over a leased line to S3 in less then 6 hours (my max. time window) and how much is it going to cost?
I'll think I'll stick with my tape loader and safety deposit box. - JasonHilton, on 10/10/2007, -2/+1Ok, so as far as I can tell, this does this only do httpdocs backups, correct? What about SQL databases?
- cbdgr, on 10/10/2007, -5/+3Ok so you have automated backups and stuff. Say your away for a day and a virus hits and deletes all your files. What happens to your automated backups?
- opensourcemaven, on 10/10/2007, -3/+1@cbdgr, unless the virus is smart enough to connect to S3 and access it, your automated backups on S3 will be fine. granted your server will be wiped and you'll have to set this all up again, but your files will be safe on S3.
- keyle, on 10/10/2007, -3/+0You gotta be good to put a bullet in my USB keys.
- coolkarthik88, on 10/10/2007, -5/+1this is very interesting. Amazon S3 is a compelling backup solution
- dezmd, on 10/10/2007, -6/+1Buried as SPAM for Amazon S3. What a load of crap.
Additionally, if you want "bulletproof" server backups it means onsite backups, multiple offsite backups across several services and/or WAN linked SANs/etc, and regular offsite physical backups, ALL of which are regularly tested for reliability. Everything else is just a finger-cross backup.
Cheers. - helpyhelperton, on 10/10/2007, -8/+2rotfl
- dknetdill, on 10/10/2007, -7/+0ooh, automatic-ness.
Browsing Digg on your phone just got easier with our enhancements to the 
© Digg Inc. 2009
— Content created and posted by Digg users is