I’m actually surprised that you don’t have tech resources at that school of yours. What about all those young minds, some of whom must be computer whizzes. 
Oh, we have plenty but the perception of the Head of School having students or staff providing tech support for his personal use would be problematic.
Hands down as individual files. So Obsidian or Finder.
Which can and do and have failed miserably. The tools available now only indicate when files ar lost in time so youcan recover them but NO, personal experiences usggests that DevonThink is very unsafe for long term retrievable data. DT also hides all the files deep in the system making them nearly impossible to recover with all the information intact if imported and indexed has other issues.
Or you can do suggested backups going back over a year and DT will show your data files all there until you actually look at one and it’s empty.
Do you mean you have never gone back?
FWIW I have recovered old files numerous times, sometimes just to compare with the current one but more often the old file is one that isn’t used or looked at for several years but when I need it I really need it. That’s why my DT data losses were so devastating. I routinely look at and use information that I collected, collated and cataloged as much as 30+ years ago. I have also migrated across multiuple computers, operating systems and formats as technology changed.Consider the length of time you will need toreference your information and how you will handle conversions as computers and file formats change.
This is a HUGE advantage over systems like DT.
Then IMO get out of DEVONThink.
Also I agree that your backup strategy is insufficient for your described needs. You need daily, weekly, monthly and yearly backups where copies are in multiple formats and multiple locations physically separated from each other for safety. And I echo that synced services are not a backup. Errors can propogate very quickly.
That’s where the mutiple and robust string of backups comes in. But if the app itself destroys the data while leaving the file and it’s metadata intact (like DT did) you will not know or find out that you lost data until it’s too late, in my case after the error had scrolled off an over 1 year backup string.
This is a great solution. I’m moving to more frequent offsite backups myself due to the nature of the data I am creating and storing.
I would say that my experience is completely different. My data loss out of DT was highly significant. Not only a large number of individual files completely gone but treatment by DT personnell that the loss was not a real issue until proved it was by others experiencing the same thing.
What would your solution be to a data system containing around 4 terabytes of critical data? Serious question.
2 Likes
Backblaze offers unlimited backup and versioning going back for a full year for around $10/month. I’m comfortable with that as my primary backup method, especially with AES 256 turned on.
If I were more paranoid and had a lot of data, and were willing to throw money at the problem, I’d be rsyncing between my local machines for redundancy, and I’d use Arq for granular control over versions using a backend like B2. If I wanted a secondary store for a mission-critical set of data I’d add Amazon Glacier Deep Archive to that stack as an absolute last resort.
2 Likes
Of those I am doing backups onto local servers and also in the process of putting in place an Amazon Glacier system.
I’ve reexamined Obsidian. But once again, I remember one of the reasons I don’t use it for short-form writing or notes (though for my research files, it would be adequate)—the Pandoc plugin does not work on iOS. 
I use my iPad frequently for writing, so this is a show-stopper for me.
I’m confused. I use my iPad for writing as well. I also use the PANDOC plug-in to do the exports. But it’s not a big deal for me to export via my mac. I have no other real way to get the data out into something else anyway from the iPad so the lack of PANDOC on iOS isn’t an issue for me.
I realize I can use the Mac, but I often travel with only my iPad. I sometimes make edits to my presentations and need to export them to Word when traveling. Accordingly, I prefer to have feature parity as much as possible between my Mac and iPad.
I have never recovered a previous version of a file as it were, I think I have the meaning right? I don’t think so anyway. I never lost anything on Mac either, though I have the back ups as I said.
All my formative years were with pen and paper and in fact often purely ‘mental’ for reasons I won’t go into, so invariably there is really only one version of anything I do on record as it were.
By the way did you ever stop missing, I might put it, the old three pane view. I have to say as time passed I became more open to your point of view on this. I keep my set up kind of similar now in fact, I think anyway.
I have recovered previous versions multiple times.
No I never did stop missing 3-pane view. I was never as efficient without it as I was with it.
1 Like
Thanks for the clarification.
1 Like
@OogieM What happened to your database was terrible and I understand why you don’t trust the software now, but you are complaining about a version of the software that hasn’t existed for several years (for most users, assuming they update regularly), and in theory what happened to you cannot happen again. In the rare circumstance of such a glitch occurring, DT would report damage long before it became irreparable. It’s of little consolation to you, but your disaster identified a weakness (DT not undertaking robust file checks) and the developers addressed it. In some ways it’s now probably more robust than e.g. storing files on Finder - how would any of us know if a file in Finder became corrupted, moved, or lost? Does MacOS even look to make sure that our files are how they appeared before e.g. a system shutdown? (I don’t know what checks Apple has in place, or how it reports an error to the user.)
For @Bmosbacker, it wasn’t clear in your comments if you’re using DT’s manual tools to their full usefulness, which I think would ease any lingering worry you have over backups. I do my backups manually too - I do a manual Time Machine backup and a manual DT backup once a week. Before I run back up, I do File > Verify and Repair Database for each of my databases, and then File > Check File Integrity. This checks that everything is where DT expects it to be, and looks how DT thinks it is meant to look (i.e. no ghost files or missing files). If everything is fine the log will report that. If something has gone weird it will report that too and you can decide whether to investigate. During day-to-day use, you might also see a log error if DT comes across something unexpected. For example, I have an indexed folder that occasionally prompts errors when I move a file via another app and DT gets confused.
I’ve been considering adding a third backup to my backup strategy, which is literally a “drag and drop” of the actual folders in my most important database on to a separate hard-drive. “Drag and drop” copies and pastes the files and folders in your database “as is”, which means they retain the folder structure they have in DT. This would backup my folder structure (and files) outside of DT just in case I couldn’t use DT ever again. As others have noted here, DT doesn’t store your files inside a program, but the filing system is opaque (to the human user) without the use of DT software. In practice what this means is that if you lost access to DT and you knew what file you needed, you could search for it in Finder, but if you just wanted to browse about, you’d be stuck. I’m not particularly worried about this because I’m not planning to leave DT, but if you copy and paste all your folders directly from DT you’d have a copy of the folder structure as you see it in DT, just for emergencies. (Please don’t actually do this drag and drop backup without research - I have not yet asked in the DT forum if there are any dangers to be aware of with this and to date I’ve only used drag and drop when I’ve wanted to move small groups of files about, not for all the contents of a database. It’s just a thing I am pondering.)
4 Likes
Thank you for the excellent input; much appreciated. I routinely perform database integrity and optimization checks. I also use DT to create an archive of my research database. Additionally, I manually export my DT research files to my Obsidian Vault, AND every six months, I upload those folders and files to my personal Google Drive. All those steps are in addition to my weekly backups to two external drives, one at home and the other at the office. Based on the manual backups and archiving I’m doing, I struggle to decide if I need to spend $100/year for Backblaze backup services. Given my current backup regimen, any thoughts regarding the value of the Backblaze backup service will be appreciated!
It’s offsite. It’s automatic. Permits you to simplify your backup processes. Closer to 3-2-1 process. It’s warranted. Cheap compared to all the guidance here (had the people charged you). IMHO.
Ok, I’ve been badgered enough. 
I just resubscribed to Backblaze.
1 Like
IF you do the verify files check regularly. But no as far as I know DT does notactually report the potential error and sicne only files that DT was “in charge of” were the ones lost the onus is on DT for the failure.
Once burned, Twice shy.
2 Likes
I do this weekly, and as I described above, I export and archive my research database consistently and even upload it periodically to Google Drive. And because I’m been harassed about this , I just resubscribed to Backblaze. I also periodically export all folders/files from DT to Obsidian. Given all of this, I believe I’m as “safe” as possible while still having all of the advantages of DT. 
1 Like
Friends don’t let friends rely on manual back ups 
3 Likes
I like that!!
20 characters…
Mac Power Users with a touch of OCD don’t tell their friends they rely on manual backups.
2 Likes