Backup Software can't Access Dataset

[MDRVa]

Morning all:

I'm new to TrueNAS and very new to Scale. I initially installed Core, thinking that was "THE" free version of TrueNAS, only to learn that Scale is now the place to be. No data onboard yet, so I moved over.

When I had Core up and running, I had a great deal of trouble getting access to a designated backup dataset from Windows machines via multiple backup programs. I've added users to TrueNAS that match my Windows creds, but that didn't seem to matter. I couldn't seem to get TrueNAS to allow anything to perform a backup -- it was like the Samba share wasn't present. But I could browse, copy and do all the other things that were expected handling data.

Eventually, fiddling around with Core, I found a "Microsoft Account" checkbox said to allow Windows authentication on newer versions of Windows . . . like magic, everything worked.

Unfortunately, this apparently useful (but hopefully unnecessary) function has been removed from Scale as noted here:

https://www.truenas.com/community/t...osoft-account-truenas-scale-22-02-2-1.102882/

That leaves me with the cool new OS up and running and no way to back up from most software packages -- seemingly because the Windows side software isn't allowed to access the dataset.

Veeam can be made to work. If you manually add the path and put in the password pair, it sees the dataset and will run a backup. But only one job, as I found out to my annoyance after spending hours getting it to work. Just one of the joys of working with free versions, I suppose. Too bad, because otherwise Veeam would be a great answer. And it might still be for that one job, which of course isn't going to suffice.

Paragon and similar will back up manually if I browse to the drive and start the process. The backup runs and completes normally. But if I so much as close the program and reopen it -- much less try to schedule a backup, I get this:

I can browse away and force it to reset, but this is completely impractical for routine, scheduled backup tasks. Other, similar experiments lead to the same basic result.

Programs like Goodsync (great file sync software) can see the TrueNAS Samba share will work, but backup software seems to continually be denied access. Goodsync is great, but it doesn't give you series of back-ups so that you can recover a 2 week old file that you didn't realize was corrupt.

At one point I sought to edit my TrueNAS user account to be sure that I had the password right -- the system accepted the re-entry, but then warned that usernames must begin with a letter or an underscore, which mine does not. But everything else works, and making a test user that followed that rule didn't seem to help. So that seems a dead end.

I don’t want to go back to Core just to get something as simple as commonplace backup software to work. I must be doing something wrong here, but there are an awful lot of possible answers. Can somebody more knowledgeable (which is likely the vast majority of the users here at present) help me understand what I think must be an access issue? How do I get Scale to allow something like Paragon to persistently see and back up to the designated dataset? And what is it that we're trying to secure that prevents this very common task from just working? Or is this some Samba problem I just don't understand yet?

Any help would be much appreciated, and I'm sure I'm going to feel a bit stupid when I see what the problem is -- lot of this stuff seems simple once you understand.

Thanks much.

 

[MDRVa]

Appreciate the folks who looked but apparently had no solution, but in this case, continued experiments eventually taught me things that may prove useful.

I had thought that the problem was one of TrueNAS security, particularly given the prior experience with the "windows credentials" check box. Turns out that probably was not the case, and that the answer instead seems to relate to the ability to specify a numeric IP address when you go to provide a destination for the backup on the TrueNAS folder or dataset.

Veeam worked if you manually supplied a path, and I had used a numeric IP address. Paragaon uses some version of the netbios name, as you can see in the pic just above. It won't allow you to manually supply the IP address. Turned out I had a copy of Acronis -- also seemingly well regarded -- that came with a WD hard drive. And perhaps by chance, the Acronis interface will let you type in an IP address, click, and then it will find and allow you to browse the dataset. And it will stick though program stops and starts, back up, and allows scheduling -- all the normal stuff.

Not sure why Paragaon can't see the netbios name on restart -- is TrueNAS withholding it somehow, and is there a switch for that? But as both Veeam and Acronis work, it doesn't seem there's anything really wrong here. Pleased, as I really do want to work with Scale. Off to learn how to do the memory mgt. hack that allows it to use more RAM for cache.

But also interested in learning, if there's a better explanation that I can provide here about why this problem occurred. For most purposes in my limited experience the netbios name just acts as a friendly alias for the traditional IP address. Apparently not in all cases.

Best . . .

 

[MDRVa]

More experiments prove me wrong yet again. The problem seems to relate to something having to do with Windows security. If i browse to the pool, I can see (and browse to) the dataset established for backup without issue. If I enter the pool and right click the backup dataset from within Windows, and if I then give "everyone" read/write/modify permissions . . . everything seems to work again.

I attempted to add my username/password on the Windows side, but that results in Windows apparently trying to assign rights on Truenas, which doesn't work. That makes sense -- but why Windows saw fit to block this in the first place is lost on me. But it does seem that this will resolve the problem. Looks as though Windows was not allowing the connection so that it could then see the netbios name on the TrueNAS side.

I reinstalled Paragon, and confirmed that it can now see, store, and back up to the dataset.

I'm sure this entire evolution has just been borne of my ignorance about file\folder\drive security -- particularly between Windows and Scale. When I get time I'll try to read further on that. But for the time being, it seems I can now make pretty much any backup software work, so I'm going to pick one and let this rest for a time . . .

Still, any educational additions here would be appreciated, and might help somebody else . . .

 

[PhilD13]

If you are trying to use the server name then it likely needs to be myservername.local as the name and not myserver as the name. Your network should resolve the myserver.local to the Truenas server ip address. As a test in the browser enter the https: // myservername.local/ui/dashboard and it should take you to the login page. If not you need to solve that first.

I would suggest reading carefully the documentation on datasets, shares and permissions. The basic flow is Pool >> dataset>>share. You access the share not the dataset. The share and the dataset both need the proper permissions set.

You will also need a user setup in Truenas that has the proper access to both the dataset and to the SMB share

The setup really depends upon what selections were made for the dataset when it was created, what selections were made for the share and if use ACL's is checked etc. and how much has been messed with since the initial setup of the dataset and share. The settings are of course located in different places in Truenas, but it will make sense why they are located where they are after awhile.

You could also take a look at Lawrence Systems on youtube. Tom has some good info on permissions.

I think the backup app will need a username and password of the truenas user that has permission to access the share you wish to backup.

 

[MrGuvernment]

"I initially installed Core, thinking that was "THE" free version of TrueNAS, only to learn that Scale is now the place to be."

They each have their own purpose, CORE is the place to be for pure stability because of how long it has been around and some other things, like memory usage for cache vs OS.

 

[MDRVa]

Thanks much for the thoughts.

@PhilD13: Appreciate the tip. I'd seen something about that before, but it hadn't quite registered. Backup issues are a bit hard to read up on in the early going, as it's often only clear from context what we're backing up and to where -- Is this client to server, or backing up the server . . . it's confusing at first. And the permission options here are many -- so I more to learn there. And I'm absolutely on the steep part of the learning curve -- the part where I often don't quite know the name of, or how to properly look for, what I'm seeking. Your approach of course makes great sense. Will try your experiments. But I did ultimately get things working, so it looks as though I can stick with Scale.

I've made several mistakes, but I think I'm beginning to understand the way datasets and shares relate in TrueNAS. Don't think I've messed up anything serious yet, but I'll check the permissions on the dataset and the share -- not sure I don't have an error there somewhere.

@ Mr.G . . . Appreciate that. Have a note to self to get the memory management hack in Scale done . . . but the new OS certainly seems stable (at least for a home use environment), and it does seem that this is the future. So for the time being, I think I'm going to stick with Scale and see if I can learn about plug ins when there's a bit more time.

 

[MrGuvernment]

For sure, in the end it is what offering suits your needs and works best for you.