Way back in March, I posted the sad story of how Mozy Online Backup completely failed me when I lost a huge amount of data to a hard drive crash. As it turned out, even though the backup client was a joy to use and worked like a charm, actually restoring my data took weeks of painful trial and error.
Even after Mozy sent me my data on DVD for free by way of apology, I still ended up losing some of it because their self-extracting restore archives don't automatically decrypt the encrypted backups and the standalone decryption tool will "decrypt" anything you throw at it, even if it's already been decrypted or you enter the wrong passphrase. As a result, I ended up double- or triple-decrypting some of my files, which essentially scrambled them. By that point I was sick of fighting with the Mozy software, so I decided to give up and cut my losses.
I did get most of my data back, though, and Mozy refunded my money and gave me a free year of service. Unfortunately, I no longer trusted them, so that year of service was ultimately worthless and I began looking for other backup solutions. In the meantime, my blog post became a forum for other Mozy users who were lulled into a false sense of security and then betrayed by Mozy's horribly broken restore functionality (which they apparently still haven't fixed).
It's now clear that, despite their good intentions, Mozy is not a reliable online backup service and should be avoided.
For the last few months, I've been using (and loving) CrashPlan. It's a little more expensive than some other online backup tools since you actually have to pay for the software itself, but it has the lovely benefit of being able to back up both to CrashPlan's servers and to your own (or your friends') computers.
As a result, I now have multiple encrypted backups of my files: one on a friend's computer, one on another of my computers, and one on CrashPlan's servers. I've tested the restore functionality, and it works like a charm. Restoring from my other computer over my local network is quick and painless. But if, for example, my house were to burn down and my friend was on vacation, I could still restore over the 'net from CrashPlan's servers.
The only real snag I've run into so far is that CrashPlan tends to be a bit of a memory hog, which is unfortunate. But hey, RAM is cheap and it's a small price to pay for knowing that my data is safe.
Rumor has it the developers are working on implementing Amazon S3 support, which will fill me with glee if it's true.
Comments
Are you still using it?
Interestingly enough, it was your post that caused me to check out Mozy and eventually sign up for their unlimited access. I haven't had a need to restore much yet, but I do agree with you that their restore system needs some work. So, now that you are using CrashPlan, are you still using Mozy as a front line of defense, or have you dropped them completely?
Re: Are you still using it?
No, I'm not using Mozy anymore. Restoring even the smallest files is a pain in the ass with Mozy, so it's not worth it for me. I should have actually tested the restore process before I began using it and recommending it. Needless to say, I won't make that mistake again.
mozy
ryan,
thanks for ripping us a new one - we deserve it. august was an ouchy month for us wrt restores. we've invested a lot of engineering effort over the last month in improving the quality of the service.
we've got over 300,000 users and so the laws of large numbers begin to kick in. i'm not talking about the data integrity - i'm talking about the application and usability. a 99% bug free app for 300,000 users means that 3,000 users run into problems where 297,000 report that everything is great. (and of course a handful of those 3,000 have blogs and as such, get to publicly flog us - which is what "helps" remind us and motivates us to improve things for the 1%.. ;-)
bugs aside, i'd say the fundamental problem with online recovery is when folks have a hard drive crash and want to restore say 50GB - this is not something you can just download in a couple of days.
i believe the "right" solution in this case is to have an external drive or USB device shipped out to our customers via FedEx. this is the only practical solution for catastrophic disk failure/recovery.
even DVD's don't scale too well when you get to disk level recovery - they only hold 4.5GB, and so for a good amount of data it's sort of like back in the old days when you had a stack of 3.5" disks to install software. yuck.
the web restore process is good for a file or two - or even good sized data sets - but it's simply not practical to recover 10's of GB via downloading, unless you want to wait a very, very long time.
anyway, that's the latest. the disk/USB restore solution we'll start to beta towards the end of the year.
in the interim it's nice to know that mozy works really well for *almost* everyone, and people love mozy - we were recently voted the #1 Best Website for 2007 by *users* in the Time magazine poll.
so thanks again for the flogging - we really are working on that last one percent!
-josh
Founder, Mozy.com
Re: mozy
I appreciate the honest and straightforward response, Josh. In spite of the technical problems, I've been consistently impressed with Mozy's customer service and willingness to own up to mistakes and try to make things right.
That said, I think you may have misunderstood the restore problems being discussed in this post. For the most part, people aren't complaining about long downloads. Obviously, when you're trying to download gigabytes of data, it's going to take a while. I think most people understand that. The problems are with the way Mozy's restore process is implemented. The Windows Explorer integration just plain doesn't work for anything other than small files, and the web restore takes ages (from hours to even days in some cases) to prepare a restore archive before it will even let you begin downloading the archive.
Even after the web restore system has prepared an archive, the download speeds tend to be bad and the connections are unreliable, which means you often have to restart a download many times before being successful. This isn't helped by the fact that the website seems to have some kind of flawed mechanism for detecting when it thinks you've finished downloading an archive, at which point it deletes the archive completely. If you hadn't actually managed to download the entire thing, you're screwed and you have to start the entire restore process over. For large restores that are broken into multiple archives (like mine was), just a single failed download out of several means you have to do the entire restore over, since you have no way of knowing which files were in which archives.
I believe you when you say you're trying to fix these problems, but please don't try to spin this as a minor technical glitch that rarely affects anyone. Mozy's online restore functionality suffers from serious design flaws that are virtually guaranteed to affect any user attempting to restore a large amount of data.
Just curious...
Can you select different data batches for local versus online back-up? I imagine I would want to store very important files online, while less critical data could be mirrored to a local server. This would ensure the survival of crucial data in case of a catastrophic event, such as a fire, without bogging down your day-to-day operations in terms of transfer bandwidth issues.
Re: Just curious...
You can't, but that would be a great feature.
Data batches (aka backup sets)
Hi GreyStork,
So far, we've purposely avoided the concept of backup sets in order to keep things simple for our end users. Every once in awhile someone asks us to put it in, so it's on the list.
Our main concern is with the added complexity, one might accidentally neglect some data and in the end, regret it.
Imagine setting up two backup sets, low and high priority, then later moving data to an external drive and forgetting to add that to the high priority set. That's the kind of thing that scares us.
As cheap as disk is these days, it seems more prudent to encourage entire copies with incremental versioning.
All that said, we'll definitely be adding the feature late this year. ;-) I'm thinking we'll hide it a bit in the UI so it doesn't become default behavior for average user.
Best regards,
~Matthew
PS: Wonko, memory footprint drops in next weeks release. :)
Off topic
Hey, I'm avoiding work at work, and so I thought I'd do a drive-by off-topic commenting. Hi! Neat to see this site is still up and running.
why?
Well dude, count me as a very satisfied Mozy customer after heeding your endorsement back in the Oregon days. Haven't tried any crazy stuff, but I did restore
S3
I'm using S3, which just got better with Panic's release of Transmit 3.6.1, which added S3 support. I can now use S3 very easily with Transmit's bucket and folder management.
I tried to use Mozy a few times, and it never quite worked right, so I gave up. After reading this post, I'm glad I did.
No title
I've been using mozy for about a month now... Fortunately, I realized it was not backing up everything it was supposed to. Tech support has replied within a reasonably amount of time, but their suggestions have been futile. I've deleted and re-installed the software to no avail. Thankfully, I haven't lost anything and needed to restore (I did test it though and after a 10 minute delay the files were indeed available *VERY SMALL RESTORE TESTED*). In the configuration program, it shows everything that I want to back up. It even gives me what seems like correct numbers regarding number of files and megabytes). The problem is, it simply refuses to back up sometimes complete directories. I'm just glad I caught the problems in time.
If you continue to use the software, be absolutely sure it's actually backing up everything you *think* it is backing up.