I’ve had two catastrophic failures in the space of a month. One was a fried motherboard and the other was the SSL bug. I am a daily archiver, so I lost almost nothing, however archives don’t cover everything. I was fortunate in both of these fatal crashes that my hard drive was unaffected. Otherwise I would have had to backdate all my user prefs, fonts, python scripts etc. to the last time I saved all that stuff, which was when I upgraded from 2023 to 2024. The changes were not significant, but if it had happened another 6 months in, they could have been since I am always modifying and updating those things. From here on in I am going to keep a go-bag off system so that if the worst happens, I can jump right back in. For those who are not archiving on a near daily basis, you are gambling.
I always made 2x boot drives for linux for each linux installation and An extra HP motherboard and PSU and GPU(could be lower specs but still compatible) and always archived daily into RAID box thats portable. Sometimes I archived every few hours if the work was very delicate.
using a Project Server makes almost all of these concerns go away. I can’t stress enough that even if you only have a single Flame, you will gain benefit. And no, it doesn’t take an advanced IT department to setup, but yes it is more involved than just installing on Flame on a Mac. @Jack should offer a service to set this up for smaller places. It greatly helps to make Flame no longer a fragile data island.
Thanks for the reminder.
Nothing like getting your fingers burnt to help remind you that everything is vulnerable.
I’ve been thinking this, especially with more and more systems running raid 0.
Maybe installing dropbox on linux and pointing to the project directory.
Or even Autodesk providing a cloud storage option that syncs the project directory?
They already have an Autodesk Drive attached to your account.
Is that a ridiculous suggestion @fredwarren
If the worse happens you can re-import the media but the setups are the most important.
It’s not so much the media, as it is the user data. My flame is heavily customized to my tastes. I have a large toolbox of user batch nodes, python scripts, expressions, as well as hot keys. It’s these that are most vulnerable. Projects and media are backed up every day. Since one of my catastrophic failures had to do with an inability to connect to my AD account, considering them as part of the solution is not an option. All I’m advocating is keeping the parts of flame that make us each unique users stored off network in a single location, so that if the worst happens, like the motherboard on your project server goes belly up on a Monday morning, you’re not scrambling to put yourself back together.
warning: I don’t believe that this is a setup that I would necessarily recommend to people but wanted to put it out there for people who may be interested.
For the last 3 or 4 years, I have run my rocky install inside a VM, using proxmox as a hypervisor. There are many reasons I find compelling to do this. But one of the nicest things is scheduled, auto-pruning backups of the VM to my NAS. I maintain about a weeks worth of rolling backups. If I ever lose a machine it is very trivial to restore the VM backup to another proxmox system and everything will remain the same. If I am about to push the boat out and try something sketchy like installing random packages, I can make a snapshot and roll back in seconds if things are broken.
This is just how I manage OS data and is independent of media
@stu as @ALan pointed out, using a Project server would externalize the project data from your workstation and ease backups and recovery. If this is too much for you, you could also use back-up applications with cron jobs to automate all this and save hours when issue arise.
Is there any documentation or instructions on how to setup a Project Server?
It’s really simple. Install RL8/9 minimal in a VM or bare metal. Configure your central authentication and NFS mounts. Then goto Flame software folder and ./INSTALL_PROJECTSERVER.
Configure your S+W how you like. Boom.
I never looked at that documentation. That is terrible and I’d say if not outright wrong, definitely overly complicated. I’ll post some clarification when I get to the office.
this is all we do to install projects server on RockyLinux minimal
./INSTALL_PROJECTSERVER --nvinst /path/to/dku/bin/system_rpm/NVIDIA-Linux-x86_64-5*
We don’t install DKU or have to do any of that manual crap that is the manual.
Then just do simple standard S+W configuration.