Before buying an expensive software for data recovery, don't forget about Photorec[0] and its dangerously powerful brother TestDisk[1].<p>It will be able to help you more often than not, and it's FOSS.<p>[0] <a href="https://www.cgsecurity.org/wiki/PhotoRec" rel="nofollow">https://www.cgsecurity.org/wiki/PhotoRec</a><p>[1] <a href="https://www.cgsecurity.org/wiki/TestDisk" rel="nofollow">https://www.cgsecurity.org/wiki/TestDisk</a>
Time Machine backups should live on Apple Filesystems. The Ext4 or btrfs in Synology's is a timebomb for Apple's proprietary data blobs. It has something to do with the filesystem attributes.<p>I asked my father to just give up on it. Just manage pictures through the file system. Photos (former iPhotos) databases also get corrupt on a Synology NAS. It's a matter of time. The strange thing is that it can go well for months on end, giving you a false sense of security.<p>Do iCloud or local drive backups or stay away from Time Machine and Apple Databases is my advice (although my father had a lot of problems as well with an "incorrectly unplugged" external HFS+ drive). I can't find the sources right now but after my father's last drama (and there have been several) I did some intensive searching and this was my conclusion.
It's so easy to think you have backups when you don't.<p>I have the following setup: my main machine is Windows and my photos are on a local NTFS drive, that is backed up (mirrored) on a Netgear NAS via rsync; I rotate the backup drives on a weekly basis.<p>Every time the backup job runs it sends an email to tell me how it went. The email is sent via gmail. This used to work well but at some point (a year ago maybe?) gmail decided this was "unsecure" and stopped forwarding the emails. Instead it sent a notice that "somebody tried to send an email and we blocked it".<p>I couldn't be bothered to fix it and accepted the gmail warning notice in lieu of the actual Netgear backup report.<p>Then eventually I upgraded the email notification system... only to find out that the backups were failing systematically.<p>Luckily nothing was lost as the main drive was fine; I was able to fix the problem and do the backups correctly.<p>But of course it could easily have gone a different way: the main drive could have failed and with empty mirrors there would have been no solution, and I would have had only myself to blame.<p>It is so easy to tell oneself that everything is a-okay when it really isn't.
I picked up a new DSLR and found out my copy of Lightroom 3 didn't support the new RAW format. Not willing to pay $125USD/year for the "latest and greatest" version of Lightroom, I found out I could use Adobe's free DNG conversion tool to just convert the fancy new format to DNG and continue using my bought-and-paid-for copy of Lightroom.<p>Also, major plus: it supports lossy conversion, which churns out files ⅓ the size of the originals, with no perceptible loss in quality. I ended up converting my entire photo catalog, saving hundreds of GBs of disk space. The tool has a CLI as well.<p><a href="https://helpx.adobe.com/camera-raw/using/adobe-dng-converter.html" rel="nofollow">https://helpx.adobe.com/camera-raw/using/adobe-dng-converter...</a>
Personally I wouldn't use that NAS for storing data until figuring out what went wrong, since <i>every single file</i> was corrupted. This is way too high an error rate and unlikely to be caused by a hard drive or network issue. It's probably an issue caused by NFS/SMB/... client/server bugs.<p>Edit: Quick googling finds many people using Macs report files are corrupted when written to a Synology NAS over SMB.
This is confusing, it sounds like the RAW files themselves are corrupt so why bring any discussion of the Lightroom catalog into it?<p>Having said that people need to take care backing up their catalog because without that single catalog file you will lose decades of photo editing work in Lightroom.
This sort of thing scares me. It's why I started running consistency checks on my important archives (like my photo library), which I keep backed up in multiple places. We tend to think that in a digital world bits are just bits and do not get corrupted — which is decidedly untrue.<p>I wrote my own consistency checker, as I wasn't happy with what was out there. I wanted it to be simple, and maintainable in the long term (>10 years horizon). See <a href="https://github.com/jwr/ccheck" rel="nofollow">https://github.com/jwr/ccheck</a> if you need something like this. I now update my checksums regularly and check for corruption.
A mirrored ZFS setup with one offline backup and one cloud backup will prevent nonsense like this.[1] If you know what you are doing, you can do a lot of stuff with ZFS even on a single disk.[2]<p>[1] <a href="https://www.backblaze.com/blog/the-3-2-1-backup-strategy/" rel="nofollow">https://www.backblaze.com/blog/the-3-2-1-backup-strategy/</a><p>[2] Forbidden Arts of ZFS | Episode 2 | Using ZFS on a single drive [<a href="https://www.youtube.com/watch?v=-wlbvt9tM-Q" rel="nofollow">https://www.youtube.com/watch?v=-wlbvt9tM-Q</a>]
Don't complicate things. I did lots of research of different NAS solutions etc. but in the end I just got a huge external USB drive and then automatic backup with backblaze for $60 per year. Works great and super cheap, plus I have offsite backup in case of fire or theft.
This is incredibly frightening. I’ve seen data rot up close, and as a former data janitor, I can sympathize.
My advice is to control everything down to the wall socket.
I considered the Synology boxes for my home NAS, but my “spider sense” was tingling, so I built around an HP MicroServer Gen8 with an LSI PCIe card, and mirrored drives. It’s on a double-conversion UPS (Smart-UPS RT). It runs FreeNAS on VMware ESXi, and the drives present as raw volumes to ZFS. No data integrity issues after ~3 years. Note the trend: clean power, good server-y hardware, matched drives in a simple RAID, properly resourced ZFS, etc.
But yeah, spinning disks are generally not a “backup.”
<i>Lightroom</i> is not a word I am comfortable reading. Not since my archive copy became uninstallable in macOS (32 bit installer) and Adobe required a monthly subscription just for me to have access to my catalog.<p>I have been made a fool of.
I don't use Apple's products. I've never used Time Machine. I have a Synology NAS though. My understanding based upon reading the article, an the comments here, is that Apple's proprietary Time Machine format is to blame here for this corruption. Can anyone here shed some light on the exact cause for this, and whether this is of any concern to a non-Apple user using a Synology NAS?
I do md5sum (shasum more recently) out of habit for anything large and at least a bit important to me. I started this about 15 years ago when I had a couple of eSATA drives and noticed corrupted files when copying. I've been checking file integrity since, and occasionally I encounter corrupted files - often enough to make me continue my checks.<p>It was an eye-opener to me to realize that data corruption is much more common than most people think.<p>On the flips-side because single (or few) bit flips often go unnoticed, people overestimate the impact of data corruption. The idea that if your image or video looks good then it must be intact is flawed. Even software binaries survive a bit of corruption pretty well.<p>I think in the long run file system based integrity checks everywhere would be great. For now, in a world with a multitude of storage technologies and file systems, shasums will have to do.
In addition to backups, error correction data is a must for things you care about. With PAR2 you can add that to any type of file: <a href="https://en.wikipedia.org/wiki/Parchive" rel="nofollow">https://en.wikipedia.org/wiki/Parchive</a>
>It's likely I messed up by not checking the "Enable data checksum for advanced data integrity" option when I created the shared folder. Still, this option is unchecked by default and it's not really clear that leaving it that way could lead to data corruption<p>What's the point of using btrfs if Synology disables one of its signature features?
I found myself, that my enthusiastic approach to photography does not utilize the CC plan on Adobe to pay ~10$/month.
What I do is to import images from SD + do a base color correction. What I found unique and useful is face recognition that Lightroom offers.
Nonetheless, I decided to move on to either open source software or software that I can buy. At the moment I'm testing On1 - but I wanted to check if somebody from HN has either own solution that wanted to share or something to recommend.
Managing large volumes of digital photos is still an unsolved problem.<p>I have tens of thousands and they are probably my most treasured possession.<p>It’s quite concerning that they are languishing on Google Photos, with a few partial backups floating around that I’m not confident in.<p>I have had a few attempts at cleaning it up but haven’t found the right software.<p>I would also like to print my favourites, which probably also amounts to thousands of pictures, but working through the sheer volume is quite intimidating.<p>Feels like the whole thing needs a better workflow around it.
This (well, worse corruption than this) is one of my worst data nightmares. Losing pictures of my children, relatives who are now deceased, etc.<p>I know it may sound like overkill, but I switched from a Synology NAS to ZFS for the checksumming specifically to avoid this fate.<p>Not a knock on Synology specifically; I amazing all are about the same in this regard. What I wish is that there were a consumer friendly low power NAS running ZFS … (?) so I don’t have to maintain a server.<p>(And a weekly sweep to AWS glacier to be sure)
>Lessons Learned<p><i>The whole ordeal was largely of my own making. I never anticipated that the simple act of copying files from one place to another could go so horribly wrong. </i><p>It isn't user's own making or user's fault at all! ( A voice is literally screaming this inside my head ) Tech shouldn't be like this. The DS218 support BTRFS which prevent certain file corruption but this is not what's happening either.<p>I have been saying ( or ranting ) about this for now more than a decade. [1] And every time the answer was there are no market for it. Or consumer or market are not willing to pay premium for high quality NAS. The Kobol NAS was my hope that a high quality, reasonably priced, and somewhat reasonably easy to setup and managed NAS would be. But they are closed down as well [2].<p>I really hope Kobol release all of the work as open source including hardware design. May be someone ( or me ) could crowdfund it once the chip shortage is over.<p>[1] <a href="https://news.ycombinator.com/item?id=10886702" rel="nofollow">https://news.ycombinator.com/item?id=10886702</a><p>[2] <a href="https://news.ycombinator.com/item?id=28299573" rel="nofollow">https://news.ycombinator.com/item?id=28299573</a>
I also have no idea how to organize photos and backups well.<p>As an engineer, I'm good doing it for enterprise stuff, but it's really challenging for personal stuff.<p>I don't trust Google products anymore, I got burned too many times when they killed their products. They work, but I just don't want to use them anymore.<p>I currently have a 100GB OneDrive subscription for $3, so that works for now, but I'm pretty close to the storage limit. I'm not a Microsoft fan, but considering it's integrated with Windows now, I assume it will stick around for long.<p>The same is probaly true with Apple and iCloud.<p>I've relocated internationally a dozen times, so a NAS doesn't work for me.<p>There are also other options for just storage like S3 or Azure storage, but the price seems to be almost the same as OneDrive.
> <i>A quick googling indicated a hard drive problem</i><p>In the vast majority of cases spontaneous data corruption happens in _transit_ due to RAM glitches.<p>All modern drives implement forward error correction on per sector basis. This allows the drive to automatically repair up to 10% of damage to any given sector... in which case correct data is returned to the requestor and the sector is either tagged for relocation or is relocated right away. In cases when the data can’t be cleanly recovered, the read request us failed.<p>That is, chances of a read request returning <i>mangled</i> data from a disk is next to absolute zero. Meaning in turn if you do see data corruption, it happened <i>before</i> this data hit the disk - i.e. it happened in transit.
I am writing a Rust CLI program for managing my backups:<p><a href="https://github.com/mleonhard/deduposaur" rel="nofollow">https://github.com/mleonhard/deduposaur</a><p>I tend to end up with multiple copies of photos and other docs: SD card, laptop, backup drives, and USB thumb drives. Deduposaur helps me to move files into my latest backup drive. It flags duplicates, modified files, and files that I previously deleted. It also checks each file's SHA-256 digest to detect corruption.<p>The program is usable but needs some usability improvements and more tests. The license is Apache 2.0. I am happy to receive problem reports, feature requests, and pull requests.
I once got a bit of a scare in that many of my raw files has weird color patterns in them. Even the backups. Turns out, I had faulty RAM and Lightroom happened to always hit the faulty parts when reading a raw image, apparently.
One thing that surprised me, if they are savvy enough to use the command line and write custom scripts, they would probably have no issue with using PhotoRec instead of buying a commercial recovery software.<p>Not that this kind of money is relevant when it's your personal files that are at stake, just a small observation plus a recommendation for whoever reads this and didn't know about it:<p><a href="https://www.cgsecurity.org/wiki/PhotoRec" rel="nofollow">https://www.cgsecurity.org/wiki/PhotoRec</a>
I use Lightroom on Windows and also Logic Pro X on Mac, and both save to my Synology. This post is terrifying me, because I don't know an easy way to validate that my Logic Pro audio projects are not corrupted on the NAS. With a photo you can just look at it. With an audio project I would need to listen to the whole thing, and when there are 32 tracks or more to be checked, that can be unmanageable.<p>For Lightroom, I keep the library on my local SSD drive and the photos on the NAS. One thing to note is that you can import photos into a Lightroom library in-place (from wherever they currently reside), or you can have Lightroom copy or move them to a location of your choice, with a few options around how to organize the files. In this case, the Lightroom software itself could well be the cause of the corruption, as it is doing the copying. I copy the files to a destination on the NAS using a drive-mapped letter, but it works with UNC paths as well. On Windows, I have ever experienced any corruption in about 8 years of using the Synology/
It looks like he’s using HFS or APFS. These filesystems still lack data checksums for data on disk, aside from any of the other issues he ran into.<p>The only sane way to deal with unique/irreplaceable data (i.e. one’s own photos) where the authoritative copy is on a mac is to make several independent copies, with several kept offline, and assume that the macOS or Adobe’s terrible house of cards that is the creative suite is going to fuck everything up at some point.<p>I say this as someone with >1000GiB in a Lightroom catalog, run on a mac.<p>I have a copy on an always-connected external disk (made with rsync && rsync -c), a copy on a ZFS NAS (made with rsync && rsync -c user@host:), a Time Machine backup that is only connected weekly, and an offsite rsync to a USB drive, with a SHASUMS file sidecar.<p>I also copy the RAWs into the LR masters photo structure when I import them, and they go first from the camera into a Syncthing folder (pre-import) that is immediately and automatically replicated onto four machines in three buildings (two of which are ZFS with autosnapshots).
I had a similar experience. Lost close to 9 yrs of photos and my wedding photos. I eventually got them back via a similar method as author. It took me a year to go through them all. After that experience I fully went to the cloud. The cause was a REFS mirror deciding to “mirror” my corrupt drive to my working drive. Great job MS. I also stopped using Microsoft for my raid stuff.
While I use Photoshop CameraRAW to edit RAW files I don’t use Lightroom or any file library management software. Everything goes into folders on my PC on a dedicated drive which is shared so our Macs can access them and also backed up to BackBlaze. Any JPEGs that I want to share with family are then copied to iCloud photo album or posted to Flickr etc.<p>I was using a WD myCloud that I retired when we moved house a couple of years ago as a secondary backup but that’s not a runner anymore. I’ve 32TB RAID6 volume on my personal office server and I’ll probably use that as a separate offsite backup.<p>We do use Time Machine with a Time Capsule that I’ve recently replaced the disk in. I don’t really trust it and also pay for iCloud Drive for most of my other data. Work also provides Google Drive which I use for data that is shared with others.<p>It’s all a bit of a mess but I think I’ve got some protection in place.
Something to be aware of the Lightroom catalog itself (which is unrelated to the OP's problem) is that it's an SQLite database. (Or it was when I ran into my problem about a decade ago bit I believe it still is.)<p>I had hardware-related flakiness with a Windows system at the time. I swapped it out for an iMac but sometime later I discovered that there was some residual corruption of my Lightroom catalog even though I had restored a backup that <i>seemed</i> to be OK. Fortunately I was able to use some SQLite tools to largely fix the catalog.<p><a href="https://bitmason.blogspot.com/2013/01/recovering-corrupt-lightroom-catalogs.html" rel="nofollow">https://bitmason.blogspot.com/2013/01/recovering-corrupt-lig...</a>
Samba (on Linux, not just Mac) has always had corruption issues for large files related to locks and filesystem caching quirks. Windows and the protocol itself has a different approach to locks than *NIX. Many have gotten more stable by disabling caching (though this hurts performance on low bandwidth / high latency fileshares).<p>Also from what I recall, Apple SMB is based on a forked version of Samba from 10 years ago. They did this when Samba went GPLv3. Not clear how well they’re keeping up with quality improvements.<p>My solution: use WebDAV on your Synology for transferring files from a Mac or Linux host. It has its own quirks but is far more reliable. Unfortunately this won’t work for time machine. But would work with “rsync -aXv —partial”.
Important backups should be stored in a variety of ways. I store my backups on multiple external HDD’s (solid state and spinning disks) in redundancy. Using multiple filesystems (ext4, and exFAT). I do use ZFS for daily/active storage but not for backups. Besides that I also store the data in AWS S3 and a duplicate in Backblaze B2. These solutions are scalable. iCloud/Google Drive is good if you only have 2TB or less.<p>Just store the raw files directly in a regular file tree structure and avoid any kind of catalog/index/db file.<p>I do find ZFS very interesting but in practical use it has been trouble working with, getting weird errors when trying to remove a vdev.
Slightly off topic:
I recently switched to Capture One for all of my photography needs. I would highly recommend it over Lightroom. Rather than subscribing to a service you can just buy it outright.<p>It has all the features of Lightroom with some extras.
Even worse because Lightroom normally does not modify the files. I am using separate storage for raw files and Lightroom database for this reason.<p>Immutable files should be the easiest case to handle reliably.
Switched to using SD cards for Time Machine-like backup (using BackInTime) plus monthly archival on big disks on Linux for many years now; requires some organization due to space restrictions, though. Distinctly remembering when iPhoto (5?) wanted me to convert/store my private photos "in the database" rather than as plain files on my old PowerBook which I refused to do however, knowing all too well this would've been a bad idea.
Nikon RAW files store a full resolution jpg version of the photo inside of the RAW file (this may be true of other RAW formats too). I had a similar problem to what's described in the article a few years ago, and for photos that were damaged which I didn't have backups I was able to get the jpg out of the RAW by running this command:<p>exiv2 -e p4 image_name.NEF<p>Apparently this jpg is what the camera shows on its backscreen when you're previewing the image/RAW file.
I haven’t had any issues with my Synology NAS yet, but I did have to stop using my iMac due to recurring _local_ data corruption errors: <a href="https://taoofmac.com/space/blog/2021/08/08/1600" rel="nofollow">https://taoofmac.com/space/blog/2021/08/08/1600</a><p>I wonder if the OP also has a Fusion drive - merely reading off a corrupt SSD cache or bad RAM would explain some of the damage.
Anyone using lightroom should also consider configuring it to always drop the XMP sidecar file with your raw files. If the DB is ever corrupted beyond repair, you'll at least have a record of your current edits per file.<p><a href="https://fstoppers.com/post-production/most-important-setting-lightroom-set-default-8366" rel="nofollow">https://fstoppers.com/post-production/most-important-setting...</a>
I keep my stuff in files only carefully categorised in directories. I use Beyond Compare to do a byte level comparison between what’s on my disk and what’s on the (one of the several disconnected) backup volumes I use. That identifies corruption issues at either side of the fence.<p>I lost about a year of family photos thanks to this sort of issue once so I am very wary of problems.
Last year all of my photos and presets had disappeared after updating the Lightroom app. That was 2+ years of edits that are just gone, lost, unrecoverable.<p>I do photography as a hobby so I never saw a need for backing up photos and I never paid for the subscription (which would include cloud storage) because I didn’t use any of the tools that came along with the subscription.
This sounds like it doesn't have anything to do with Lightroom, but is instead file corruption caused by some kind of bug.<p>Lightroom catalogs also don't contain any images, I think a lot of people are confused and think it does.
Based on this, and on my own experiences, it looks like it's time to move on from Synology. But... to what?<p>I've had a DS412+ since, well, 2012. I've dutifully replaced drives as they went bad, run disk scrubs, file system scrubs, and (in the last nine years) had to occasionally run time machine restores from it. It worked like a champ each time, including "restoring" from an old computer to a brand new one. However, in the last couple of years, we couldn't get my husband's machine to back up or restore successfully, and now out of our three macs, only one is able to successfully back up at this point, and (looking at this) I'm skeptical that it's actually working.
My bet is the problem is related to the USB transfer. I've read about similar problems migrating data to Synology via USB. Don't remember where I read that, though...
Gaaaah.<p>This is mostly not about Lightroom, or photos, or any of those specifics. This is about (a) a failure in a migration followed by (b) *failure of the user to verify the migration before deleting the original copies*:<p>>To make matters worse, I deleted the photos from the USB hard drive, without verifying the new catalog.<p>That line is, not coincidentally, where I stopped reading.
What I find work best for storing data is a regular USB drive. With no software like Lightroom, no NAS. Just a disc, you move data with a regular file manager.<p>You don't get all the cool features, but process is pretty safe and you don't worry about hackers because it is offline most of the time.
Apple wants you to buy a whole new MacBook for the extra storage, and only use their icloud for backup. Apple intentionally has bad implementations of network filesystems. They know it, but as per usual, they know their customers won't do anything about it.
I use lightroom and i have over 3TB of photos. This is my current workflow. There's no reason why it would not work on a Mac<p>1. Lightroom runs on Windows 10 system<p>2. I add new images onto it from one set of cards. The cards are pulled out and sat on a desk. A previous set of cards is added back to the camera and wiped. At this point there are images in the Lightroom and images on the the cards.<p>3. Every night, windows box rsyncs the Lightroom catalog and raw data onto a Linux box running in a RAID-1<p>4. Every night, after the rsync is completed Linux box triggers a backup of the rsync'ed catalog and files into S3 compatible remote bucket.<p>This means that I have all images and catalog stored at least twice locally and once remotely. I have also validated that I can recover from a failure of Windows disk, Linux server and remote backup.