Digital Media | Spotlight: Photography | Inside Lightroom | Blogs
The Simplest Photo Storage Solution
Starting out on New Year's Day, I've been writing about storage here on the Inside Lightroom and evidently it's hit a nerve as my fellow blogger Micheal Clark has also chimed in on storage solutions. What with all these posts about storage, it might start feeling like the all-storage channel here. On the other hand, it's an important issue. You need to make sure that you're photos are safe after you take them and you need a plan to get there. As part of my New Year's Storage Resolution, I've been looking at lots of different ways to tackle the problem. And, after exploring all the alternatives, the simplest strategy is exactly the one that Michael talked about yesterday: A simple array of cheap mirrored drives.
To prove this setup out for myself, I built out a version of it at home last week sized for my current needs. I went out and bought three 750GB internal Seagate SATA drives, and three 750GB Seagate FreeAgent USB/eSATA external drives. The eSATA connection was important to me as I have a PCI card in my desktop that provides eSATA connections and it's the fastest way to hook drives up to your computer, blowing away USB and FireWire. As a bonus, I was able to pick up the USB/eSATA drives a bit cheaper than FireWire equipped drives. As of January 2008, the price for all these drives is a bit under $1300. Not bad for 2.25TB of mirrored storage. If you don't already have an eSATA card, you'd either want to pick up the FireWire versions of the drives, or spend $200 or so on a good eSATA card.
The most important part of the puzzle aren't the drives, however. It's in the software that you use to automate the backups of your drives between each other. For this setup on a Mac, I ended up going with Carbon Copy Cloner to synchronize data between the internal and external drives. Once a day, everything gets sync'd up. And, if my system is asleep when the time to sync comes around, Carbon Copy Cloner will kick in the next time I wake it up and make sure that everything is taken care of. Automating the backups is the key component.
As far as organizing photos on disk, I've got the 70,000 photos contained in my Lightroom libraries up through 2005 on the first disk dedicated to photographs (Photos Alpha), and my 2006 and 2007 libraries on the second (Photos Bravo). Depending on how my shooting volume goes for 2008, all of those photos will probably go onto the second drive as well, but I do have space for a fourth drive (Photos Charlie) in the form of another pair of drives. At that point, I'll have to think about upgrading to higher capacity hard drives as I'll be out of internal drive bays.
So, what about that all important third backup? The one that goes offsite? For these backups, I going to use a stack of my smaller 200GB and 300GB FireWire drives—the ones that used to be my primary drives a few years back. Once the data is dup'd over to them, probably organized by year, they'll go down the street to my bank's safety deposit box. And I'll start sleeping much better at night again.
To finish up, here in no particular order are the answers to a few questions that you might ask about this setup:
Why 750GB? After all, with 1TB drives out there, wouldn't it make sense to maximize? Well, it probably would. But currently, the 750GB drives have a better bang for the buck than the 1TB drives that I priced. That won't be true for too much longer, probably, but I was looking for a balance between capacity and budget. After all, there's some new lenses I want to buy.
What about Super Duper? Super Duper is a great tool, but at the time of this writing, it's still not in form for running on Leopard. It's coming soon and I'll probably use it when it's released as I really like it.
What about a tool like Carbon Copy Cloner on Windows? There's got to be something on Windows that does the same kind of thing. If you know of one, please leave a suggestion in the comments.
Why Seagate? I've used all sorts of drives in the past and while some studies show that there's no real correlation between brands and failure rates, I've definitely had good luck with Seagates, so that's what I bought. But, its important to note that hard drives from all makers fail. This is why you need backups in the first place.
What about using a Drobo or a NAS? I've looked at these solutions and may still invest in one of them for another layer of backup. But, for the time being, I was going for cheap and simple while staying robust. I really dig the idea of the Drobo, but really would like to see it sporting a FireWire or, better yet, an eSATA connection. Also, I've seen that the Drobo now has the Droboshare which lets you connect up 2 Drobos to your network, but the catch there is that the connection between Drobo and Droboshare is still USB. Darn it.
Come on, don't you really want a big RAID setup? Not really. As I've said before, RAID is not a backup solution. Also, RAID is expensive. An XServe RAID with the roughly the same capacity the setup that I've put together is over $8000. And, you'd need another layer for a second copy backup. Sure, with RAID 5, I could access that data more quickly, but disk I/O speed isn't really the biggest problem in my day-to-day life right now.
What about bit rot? I have to tell you, bit rot scares me. Say one of my RAW files on my internal drive gets corrupted. That corruption will be faithfully copied out to my backups. The only defense I currently have against it is the offsite backups. I'm actually considering making a second set of offsite backups of all my RAW files that never gets written to again to help guard against it.
What about ZFS? I'm so ready for ZFS on the Mac. And they're working on it. As soon as it's ready for prime time, I'll most likely start taking advantage of it. But, I'm not going to set up a second server for it to run at this point. That's a cost outlay in both hardware and energy costs that I'd rather put into more camera gear.
Ok, that's about it for storage and me for a while. I'll change gears next week to something else. But, I'll be keeping an eye on the storage space as it develops.
It's funny, that's the exact same setup I have! Except I did it earlier so went with 500GB Seagate Barracuda (internal) and FreeAgent (external) drives as they had a better price per GB at the time.
"What about a tool like Carbon Copy Cloner on Windows?"
Acronis True Image 11:http://www.acronis.com/
I have tried many other solutions, this one is "best of breed".
... additionally I would suggest a folder synchronization software,Allway Sync: http://www.allwaysync.com/ and Microsoft's SyncToy (v2 currently in Beta): http://www.microsoft.com/downloads/details.aspx?familyid=c26efa36-98e0-4ee9-a7c5-98d0592d8c52&displaylang=en
I also recommend SyncToy, which was developed with digital photos in mind, I believe. I've been using it for six months now and it works like a charm.
I'm all for disk redundancy, but off site backup is needed if your archive has sentimental or monetary value. I use Mozy now. Now I know you must have a tremendous amount of data to backup. That being said I had to backup about 170GB and that went across the line in about three weeks. Every time I shoot Mozy backups up the incremental data. Also one of the biggest benefits for me is limiting the bandwidth. I don't pay for any increased bandwidth on my cable modem so I set Mozy to throttle its bandwidth while I'm at home, usually 6PM-1AM every day. I would highly recommenced supplementing your backup strategy with an online backup.
James
James I'm with you 100%. We use an offsite backup solution to store files remotely and for synching both our inhouse systems and online storage systems we use Synchbackse http://www.2brightsparks.com/syncback/sbse-features.html
For incremental backups of the offsite systems we use rsynch on Linux. The fact that our building could burn down with all our working files and inhouse drives ... and a pc/mac with an internet connection could pull back in all those files from the offsite backups for a realtively easy rebuild - allows us to sleep easy at night.
Despite the challenges of uploading so much data over the web, we placed a full T1 from Speakeasy on our studio for only $400/mo. http://www.speakeasy.net/promos/fatpipe/ and mitigated that issue quite a bit.
I think that is really too expensive this kind of backup... but no one considered my previous comment on this?
with RAID 1 everything is mirrored in realtime without buying any other software (that could mess up everything if crashed)...
James,
I use Mozy too. If you haven't run into this yet, let me offer a heads-up. If the backup ever runs and the device you are backing up is offline, Mozy will delete all of your files from their site, requiring you to go through your multi-week backup to get them all back up there. I now have automatic backups suspended and run it manually once a week at minimum. I also run it manually immediately after importing important photos such as a wedding shoot (after double-checking that the drive is mounted).
In terms of Super Duper, they say it's not fully Leopard compatible yet, but I have to say, I've been using it with no problems (and have even used the backups). I realize you've got a very sensitive/important set of data here, so it may not be worth the chance, but figured I'd mention it.
"Also, RAID is expensive. An XServe RAID with the roughly the same capacity the setup that I've put together is over $8000."
It's XServe which is expensive, not RAID. You can get yourself a budget PC, install linux and do software RAID for a couple hundred bucks plus drives. (Or, if we weren't trying to do it on the cheap, drop a few hundred more on a good hardware RAID controller.)
James -
Very well said - some great info here and as always the best way of learning is from other peoples experiences. Thanks for letting me chime in on my blog post as well - great to see what other folks are doing in this department.
Cheers, Michael Clark
I presume the reason you don't use a photo-hosting service like smugmug or zenfolio for effective backup is the upload times & dependence on a third-party? Both supposedly have unlimited storage for around $100/year, thrice-backed-up to geographically separated storage. There are now uploading plugins for most library-management tools, like LightRoom, so it's pretty simple.
John.
Backups are one thing but in a related twist, what about synchronizing Lightroom edits on two Macs?
Some editing work may be done on a laptop, while edits on other images may be made on the desktop.
It doesn't seem to me that CCC will handle this (in other words each machine may contain relevant edits unknown to the other)
Is there a something out there that will update two Macs with the combined data to reflect both the edits made on one machine in the morning, and on the other in the afternoon?
Or has a solution been written up before in this blog, and I've missed it?
Any recommendations for a "good eSATA card" for a Mac Pro?
Great write up Duncan. I really need to get a safety deposit box. Otherwise, I've already implemented most of what you wrote. Been doing it for a while now. (Actually, I haven't finished bringing all my old photos into LR libraries, what a chore!)
Thanks to all that chimed in with Windows solutions. That's way cool of you.
James R Marcus: I agree that offsite backup is essential. As I've posted on this blog before, I believe that both online and offsite are good options. I'm ending up at the point where I'll be using both, especially since an online backup can be accessible from wherever you happen to be if you really need a file in a crunch.
Mauro: Indeed, setting up another server is a valid strategy. After all, it's common enough to have extra machines hanging about. But, for me I didn't want to go down that path as it takes more power to run multiple machines than I want to expend right now. Also, administering another machine isn't high on my list of priorities. I've been weening myself out of the sysadmin game for the last couple of years.
Rick: Sure, an XServe RAID is expensive. But even picking up a budget PC adds expense and system administration time. If you're up for admining another machine, sure, doing something like that works well.
To all that talked about Mozy: Thanks for that. I've been currently looking at both them and S3. I'm currently leaning towards S3, but haven't made a final determination.
Michael Clark: Of course! We're all adding to the discussion and learning from it.
Les: Synchronizing edits is way out of what SuperDuper/CCC will do. I always follow a workflow where if I do work on a laptop, it migrates to my desktop and then that's the master copy. Keeping more than one library going at a time with the same photographs is a headache x3.
Jeremey Pinnix: I'm using FirmTek's eSATA cards in both my desktop and my MacBook Pro. They get the job done, tho I haven't looked at any others. I went with FirmTek based on recommendations that were sound at the time, but that shouldn't rule out any other cards.
Thanks all for the great comments!
Regarding S3 vs Mozy: I am using both. Mozy is definitely a cheap solution. However, once I needed to have a single 50k file restored and it took them 3 days to restore it. I don't know about you, but that was basically useless to me. So I am now backing all essential files up to S3. It's going to end up costing me $30-40/month, but the piece of mind is worth it.
Jermey: Yikes. That's not a good thing at all. Thanks for the info on that as I'm currently evaling several services for my own use and for reporting back on this blog in a week or so.
James, thanks for this post. I identify with your approach - I have too much to worry about to pour a lot of maintenance into a storage solution.
One quick question for all your readers, and apologies for the PC focus: can anyone recommend a good eSATA PCMCIA card? I've researched several cards and it seems in addition to the PCMCIA bandwidth bottleneck, some cards run software that further limit the bandwidth. Other cards will divide the available bandwidth by two (amongst the two ports) even if only one eSATA drive is plugged in. Has anyone done the research on which eSATA PCMCIA card offers the highest bandwidth?
Thanks in advance!
I purchased the Burly 4-drive enclosure, four 500GB WD drives and the Sonnet X4P PCI-X card for my powermac dual 2.5gb G5. I am trying to set this up with 2 pairs of the drives striped (Raid 0) and then I want to mirror those two pairs (Raid 1) for redundancy. Can anyone explain to me how to do this through OSX (if possible)? My wife is a photographer and I want to provide her with a safe storage solution between hard back-ups (ie: DVD-Roms).
My wife wanted the wiebetech storage to do raid 5 but I went this route thinking I could save some money and accomplish the same thing. And the wiebetech would have been firewire or USB and I thought I wanted eSata for faster write times. I am hoping that I don't regret my decision.
All advice is very appreciated.
DS