I work in a niche inside a niche. I deal with terabytes of storage, massive servers, a variety of storage tech, and I’ve been in interested in computers in general for… Around 40 years. (Yeah, I’m old.)
I have my own single person company and have worked in 40+ US states, done assignments in the UK, Norway.
AMA.
deleted by creator
I normally don’t get into storage at that level – most of the storage solutions I use are enterprise-grade disk from Hitachi/Dell/EMC/IBM, etc. I’ve looked at Gluster, but never really tried implementing it. Ceph seems powerful, but more complicated than I’m willing to get into for my personal projects.
Once you cross a couple terabytes, I start moving everything to tape or (more recently) cloud storage where the day-to-day management is someone else’s problem.
In your opinion what would be the best archival format for storing photos and videos of the family. Without relying on a ZFS server running for 20 plus years, but a “hard” copy like Blueray M etc
So no suggestions? dealing with data yourself you must have your own best storage go to? no?
Honestly, I’m “storage agnostic” – in my office I have Hard drives, SSDs, NAS, servers with various types of RAID, Linux boxes with disks in LVM, magneto optical platters, and various tapes.
It’s less about the media and more about the process. As I described elsewhere, I have a large NAS, an onsite copy, and an offsite copy on tapes. It’s the process of keeping offsite copies, regularly updating them, and verifying the copies that protects me, not some sticker on a box that says “100 YEAR STORAGE LIFE” from a company that might not exist next month.
Yeah, just curious. i have heard tape is a decent option or Archival DVD. Running a server and backups what I do now, but it is not really a way to pass on family data like you could with a photo album. Especially when they are less tech savvy family.
Every media is subject to failure. It’s the process that protects.
If you’re keeping something for your family, consider putting it online on a sharable cloud storage system, or using software that distributes the data to everyone’s computer (BitTorrent / Resilio Sync / DropBox, etc.)
If you want something physical, I’d get a ‘tough’ or ‘high endurance’ USB stick or SD card, and keep updating it quarterly. Flash doesn’t have a great reputation for longevity/durability, so I’d wipe the USB stick clean with zeros then re-write everything with each update.
Thanks!!!
Archive.org, one of my favourite websites, has been targetted by litigation.
What do you feel will happen to this website in the years and decades ahead? And what will happen to digital archiving should the site be eliminated from the web?
Can you get in touch with me? I work in archives, in IT, and have a nasty situation I’m looking for advice on from someone with experience in exactly this. Can we dm? Not sure how that works here.
deleted by creator
Awesome, I appreciate the offer!
Sadly, no… My niche is so very, very small that it’s unlikely I can help your specific situation. It’s also a self-preservation thing – giving professional advice for free without contracts in place is a liability issue.
Those two reasons pretty much cancel each other out, but alright.
What is the best filesystem for archive ? BTRFS, ZFS, ReFS?
How will quantum computer affect your field ?
What is the 2023 bottleneck ?
IMHO , it’s permanent storage . I remember in 2005 up to 2010, we all wanted the fastest CPU , more GHz , more cores ,etc. the industry gives us that . Then during the last decade , we all wanted a more powerful GPU, more core ,more memory ,more Mhz. We also craved faster internet connection ,now we have optical fiber with 3.5Gb/sec in home for 120$/month .(impossible to think in 2014)
Nowadays , I have the impression that permanent storage is lagging behind . With the new medias being in 8K, video games storing average 100Gb and what not, we regularly move around dozens of Gb, even as casual users .
I’m only familiar with ZFS, but only in my lab, not in production… ZFS is great because it can self-heal files / re-allocate blocks. I tried it on SMR drives, and it’s terrible, I advise against it. :)
ZFS is very good, but OFFSITE, TESTED BACKUPS are critical. There’s ‘reliable’ storage (storage that can deal with a failure) and then there’s backups. All the parity in the world won’t save your data from a fire.
In my small office, I have about 100TB of data that’s important to me, so I have a local copy, a backup in my office, and a stack of tapes at home about 1km away. Anything that affects both locations is outside my threat model, as I’ll have bigger issues.
Always be ready for the TornaDoS attack.
The irony being, just last week, the city I live in just had its first Tornado warning in nearly 50 years.
What are your feelings on tape backups?
Most people in the industries I’ve worked in (mainly SMB, MSP… Sysadnin roles), seem to think that tape is an archaic method of doing backups, and anyone using tape is living in the past.
Additionally, for archival/backup software, what’s the go to for you? Both paid and Foss, if you have options for both, I’d like to hear it. What makes it the go to software?
Thanks.
Tape is awesome. Relatively inexpensive at scale, huge storage volumes, consumes almost no power compared to what it stores. But it has its time and place. That place is archival and long-term offsite backups that are very infrequently accessed. People aren’t using it for what it’s best at doing.
The backup/archive software I use for work is enterprise grade - Tivoli Storage Manager a.k.a. Spectrum Protect. In my office, I use Time Machine on the Macs, and simply ‘tar’ on Linux to back up specific important directories. Windows machines are backed up by their owners with various tools that I don’t tend to concern myself with.
For the enterprise stuff, what makes it great is that it gives you a huge amount of control and flexibility and storage options. I love the idea of TSM/SP’s ‘incremental forever’ backup methodology. It means you can roll back to any backup at any point in time, as long as you’re storing enough historical versions of the files. The device support is also amazing, and I’ve built systems that can scale to be petabytes large with it.
For my office, I just use what I know is built in and reliable. I know every Linux system has tar, and every Mac has Time Machine. For my NAS device, I make copies of it with rsync to a USB-SATA enclosure with 5 drives, usually every 90 days or so, less if I’ve made a lot of changes.
