That shouldn’t be the case…
Well, in my case I found screens that should be virgin containing data from previous experiments. A 128 GB SSD only takes 40 minutes or so to format so I don’t have to deal with potential landmines.
My ability to fail doesn’t need any help.
I’m quite curious about the thought process that leads you here. As with the iSCSI thing, I think it’s hiding some incorrect underlying assumptions which are going to make your use of TrueNAS much more frustrating. What about this seems “clunky”? Because one share = one dataset is an extremely common–probably the most common–arrangement.
1 Like
Wiping a SSD should be instant…
Either you’re doing things strangely, or the ghost of Captain Murphy keeps you on close watch.
Refreshing and clearing cache often helps with such anomalies.
I’m aware of having to unlearn dead-ends is a problem when your knowledge base is 35 years old (I’m 73). Experience is a two-edged sword.
Right now I have three datasets and three users pointed at various dataset combinations. I’m puzzling what to do in “Sharing”, just point sharing at “pool” name/folder containing the three users? Do I need to add something to the first line “Path”?
I see a few more clean-slate do-overs in the future.
Clearing cache didn’t occur to me. Good thought.
Wiping the directory is a 3 second operation. An actual full disk format takes longer.
If I’m rewriting from the same installer to the exact same sectors of the SSD drive with only the directory wiped, unwanted artifacts seem to me to be potential termites in the woodwork.
I personally avoid directory-only wipes whenever possible. I’ve experienced disk sectors overwritten several times, causing all sorts of unwanted gremlin-like sabotage.
For anyone getting frustrated because old data is messing up attempts to enter new data. I formatted my 128 SSD and cleared the browser cache.
I went to create a new pool on a fresh TrueNAS install, and up popped the pool from my previous build attempt.
I’m thinking of digging out a different 128 GB SSD, a normal format only flips a small number of bits, nowhere near all.
(I seem to remember a high-security military wipe requires 7 full format passes.)
Of course it did–the pool doesn’t disappear just because you nuke your boot device.
But I hadn’t pointed the fresh system at the pool I hadn’t even clicked “ADD”.
I’ve never seen, in several reloads in this session, the system latch onto the storage array independantly.
I also have to assume the pool now has some switched bits just to make life interesting.
I have experienced a ghost in the machine from previous data a number of times. It was much worse back in the 5-1/4" and 3-1/2" floppy days.
I’ll just reformat every 45 minutes or so this evening and start again tomarrow.
Instead of formatting and reloading the TrueNAS system onto a 128 GB SSD, I switched to a 32 GB SD card and an SD to SATA adapter, with the TrueNAS system installed on the SD card.
(No noticeable speed difference so far, and the SD to SATA adapter costs about the same as a 128 GB SSD.)
Then I use Windows and balenaEtcher (the only time I ever use Windows; my go-to is Linux) to clone SD #1 onto three other 32 GB SDs. No more formatting “ghosts”; I reclone SDs as needed on a third computer.
(Will a fresh clone leave “ghost-code” like full formatting? Unknown territory for me.)
All my computers have six system selections with Windows 11 as one choice, SSD/HDD hot swap, and SD/mini-SD card slots as other options. Or, any combination thereof as needed.
I’m hardware-oriented, not software-oriented (dyslexia); I’m wondering if this multiple datasets quest is worth pursuing. I already have a one-user-fits-all TrueNAS server that works just fine.
However, I’m retired, and I have time to burn.
This is going to wear out and break. There won’t be noticeable speed difference. It’s a bad idea because it’s the wrong hardware.
This isn’t accomplishing anything. I’m a bit shocked that permissions are so baffling if Linux is your OS of choice.
You don’t strictly even need multiple data sets. You just need to create your users and set the correct permissions on your different storage areas corresponding to each user.
The least confusing way is to do this with 3 datasets, because this presents you with an easy GUI interface to set the permissions accordingly.
Permissions, in the form of ACLs, are just a list of what user(s) can access each dataset.
I first pursued one dataset with different users with different permission parameters. But how to give two users their own password protected, size limited folders as well as access to the “public” server that less privilaged people can explore has been a flop.
I started out with an Apple II with dual 5-1/4" floppy drives and thought a pitiful Winchester HDD was fantastac…These past decades have left my head full of obsolete terminology and methodology…Unlearning/relearning with sever dyslexia is a bitch. But at 73, I won’t have to put up with this shit too much longer.
I sort of thought ACLs were the key. I just need to understand the lock.
To be fair, I and lots of other people started out on Apple IIs with only 1 floppy drive and no hard drive too. SMB sharing and permissions have been pretty mainstream for 25-30 years at this point.
So first of all…we’re several years in to anonymous or “guest” access just working. It’s insecure and turned off by default in windows, for example. So be aware that you really want users connecting with a username/password.
Secondly:
with SMB authentication occurs at the USER level. You’re not password-protecting folders - and indeed you can’t. The server asks the user for a username and password upon connection. That username and password is then used for all communication until the user would do something like log out of windows.
You tell truenas what users are authorized to access each share/dataset
So this is a fit if people are using their own computers to access, but doesn’t really work for a shared PC that people aren’t logging in/out of between different users.
These limitations sent me on the multiple datasets track.
I pretty sure TrueNAS can do what I want.
What would be great from my standpoint is each user logging into the same storage array via different Truenas systems. Having three TrueNAS system drives with different network addresses would be easy. Having all use the same RAID storage array (pool?), not so much. My network isn’t accessible from outside, by choice.
My work history is in R&D, dream up new and better ways to fuck things up.
Maybe I missed something, it’s a bit late and I’m quite tired…
I’m not sure why you’d want/need three separate systems with different network addresses.
Use permissions/ACLs to set permissions for each user/group on the folders you want each user/group to access, on a single TrueNAS system.
What would this accomplish?
So is having 1 TrueNAS system with 3 different network addresses. But it doesn’t solve anything.
The three systems thing is approaching the issue from an entirely different direction habit I developed over the decades. If my whack-job thinking makes your brain hurt, just ignore my tangents.
No need for any self-deprecation. You just seem to be overcomplicating something pretty simple.
I don’t feel like scrolling up, so I’m going to draw some assumptions from memory of what I skimmed earlier…
Figure out your users:
bob, jane, bill, mike, bozos
Figure out your groups and add users:
admins: bob, bill, mike
legal: jane
executive: bozos
Figure out your desired folder structure:
share
share/admins
share/executive
share/legal
share/home/bill
share/home/bob
share/home/jane
Set the permissions for each folder with the appropriate users/groups.
Everyone can access the same share/ with their username/password at the same network address on the same TrueNAS server. Users/groups can only access the folders/files you’ve given them access to.
A bit oversimplified, but hopefully you get the idea. It might help to just describe what ‘end-state’ your looking for, as far as users/groups/folders.
From your original post, you don’t really set user/pass per folder, you set user/pass per user and then use permissions on the filesystem to control what that user can access within the share.
I don’t tend to depreciate myself, I often see things from several different angles at once.
I was born with eplepsy, sever dyslexia, and autisim (called Kanner’s Syndrome back then). I spent my entire 18 school years in “Special Education” AKA: Retard classes.
I will try what you wrote, it certainly looks impressive!