Nov 10 2009
Twin by App4Mac Review - $70 in the toilet.
I’m an avid user of Time Machine to keep all of my media (most importantly, digital pictures dating back 15 years, and scanned images dating back nearly 20 years) backed up and safe, but a few months ago I started to consider the possible issues of having a backup sitting right next to the source.
In a situation of hardware failure on our iMac, the Time Machine backup is indispensable - simply fix the hard drive, and restore.
However, in a situation of double-hardware failure (unlikely, but possible), corruption (again, unlikely, but possible), or (hopefully never) catastrophic loss of our household contents, a local-based backup is going to prove rather useless. I needed an off-site backup solution to complement the local backups. Risking nearly two decades of memories simply calls for more than a lackadaisical approach to backing up.
My current webhost provides me with an unbelievable amount of disk space and the logical solution was to simply backup to this otherwise underutilized space, but simply compressing everything up into a huge archive and FTP’ing the files to the server seemed like a pain in the ass, and since it wasn’t easily automated, it was prone to being forgotten and as the months pass by could still result in file loss.
Enter “Twin”, by App4Mac. It seemed perfect, and in testing the feature-limited free version, it also seemed to actually work. It was straightforward to setup, was highly automated, and was able to utilize my wasted server space as an off-site backup destination. It was exactly what I wanted.
I was a little put off by the nearly $80 price tag, but I tweeted the developer and was sent back a “20% off” discount code which took a bit of the sting away from the purchase, so I took the dive.
The problem is that although it seemed to handle the small amount of files in the demo version, once I was registered and I set Twin loose with my 60 gig iPhoto library, things started going horribly wrong. The first failure was a result of me pausing the backup (using the server applications pause feature) since I needed the bandwidth for something else, except when I went to continue the backup the server crashed. 30 Gigs of upload wasted.
The second failure happened after about 40 gigs of the backup had completed. Starting from scratch.
The third failure happened after about 50 gigs of the backup had completed. Once again, starting from scratch.
You see, aside from the fact that the program doesn’t seem to be able to reliably handle large backups (as is evident from the crashes), it also can’t recover gracefully from said crashes - if a backup is not completed to 100% (a process that can take a very long time with a slow DSL upstream connection handling 60 gigs of upstream content), the program simply orphans everything that was backed up to the point of the failure and wants to start all over again when you restart.
Ridiculous you might say, but after discussing the problem with the developer, confirmed. Apparently no interim index of check files are written during the backup, and if it fails to complete for any reason whatsoever, you loose everything. Checking my server I can confirm that my last backup attempt successfully transferred just over 50 gigs (spanning the period of numerous days worth of saturated upstream transfer) and the files are there, but there is absolutely no way to re-associate those files to Twin and make it continue where it left off. Even if I attempted to manually handle the files that it did manage to backup it would prove infuriating since Twin aggravates files into batches, ZIP’s them, splits then as required, and (I understand) renames the source files to fit it’s own backup schema.
There is a feature to “recover an orphaned backup”, however again, it only works if the backup in question was 100% completed (at least once) to begin with, and then the Server app somehow lost it’s configuration of such.
So, after wasting over 120 gigs of bandwidth, I’ve given up. The developer did last respond saying a future version will address this issue, but when I inquired as to the anticipated release date of the new version, I received no response. I’m considering asking for a refund and looking for an alternative backup solution that does what I need, but I can’t seem to find anything else that handles things quite the same way Twin does - this is of course why I bought it to begin with.
The only good part of the story is that I have an ISP that offers me a generous 250gb/Month allowance so these failures aren’t going to cost me an arm and a leg in overage charges, but I still can’t help but feel a bit ripped off.
You didn’t really ask for advice, but I also do a multiple-tier backup. Time Machine in the house, and BackBlaze for off-site.
I have nothing but good things to say about BackBlaze. You should realize that if you hit their web page from a PC, it only tells you they have a free PC demo, causing me to send them an email saying “but you said you had Mac support!” and they said “it’ll show up when you hit the page from a Mac!” and it did. That led to me exchanging a few emails with their tech support, and they were remarkably competent. I suppose it says a lot about the software industry that it’s worth mentioning that, but that’s a rant for another day.
Now, it’s not fast. It took me something like 3 weeks of letting it run every single night before it got my entire 100GB or so of data completely backed up. You can throttle the network usage and let it run 24-7, but I just turned it off when I was at home and gaming. Once it got the initial backup done, I just let it run at about half speed all the time, and it’s always up to date.
It was perfectly happy being started and stopped multiple times a day - it would just pick up where it left off. It understands Mac bundle files, so it knows it can back up the iPhoto “library” one photo at a time, for example.
You can restore a few files at a time from their web interface, you can get them to email you a zip file with a lot of files (both free), or if you have a really catastrophic failure, they’ll mail you your entire backup on DVD for $99, or for $189, they’ll mail you a USB hard drive with all your files on it.
And it’s $5 a month or $50 a year (for each machine, although I think I read something at some point about multiple machine bundles). That’s worth it for me to have one less thing to worry about.
Have you considered rsync?
If your web host offers SSH access to your data, you can use rsync to do your backups.
Its a command line utility that is really smart about everything. Ie, it will only backup the CHANGES (incremental). So you will always have a full backup on the remote side, but if you make a small change to a file, it will only upload the different, not the entire file.
Check out this page:
http://www.bombich.com/mactips/rsync.html
Unfortunately the hosting account I have at 1&1 doesn’t offer SSH - the next account level up does, but the cost goes up quite a lot and apparently they can’t migrate anything on my current account to the new one, meaning I’d have to do a complete backup and migration - in other words, a PITA.
For the record and anyone finding this via Google, Twin is STILL not working right - again it will not recover gracefully from an interrupted transfer.
I had completed over 30 gigs of backup on my latest attempt only to have our iMac freeze up courtesy of Spore (Freakin Electronic Arts junk software!) and when I was forced to reset, again Twin acted as if nothing was on the server and wanted to start all over again.
@Mark: Since you mention 1&1, check their terms and conditions. For my package, they state
“You shall not use the Web Site Space or Your Services in any way which may result in an excessive load on the 1&1 Equipment, including but not limited to installing or running web proxies, using your allotted space as online backup or storage, or mirroring mass downloads.”
Di you find any other solution? I am using “Arq” so far - works great but I would love to skip the Amazon AWS S3 storage costs …
Best,
Steve
Since the original version of Twin that I purchased the software has gone through many revisions but none of them still appear able to recover from a catastrophic (read as, power failure or the software crashes) termination of the backup process. I tested the current version by starting a backup and then about an hour in manually killing the process in Activity Monitor and upon startup it had no idea there was any data backed up on the server (despite me being able to see it there via FTP) and as such simply started over again. With a small backup not a big deal I guess, but when we’re talking 60 gigs, a failure half way through is massive and time consuming.
I’m currently in the process of attempting another full backup. I have a UPS on the system in question now to hopefully avoid any unexpected (however brief) power interruptions that could screw with it, but if the software craps out or something odd happens (as always seems to when I commit to these large backups) it could all be a waste again.