Batter Pudding

Ideally a second Unraid with Urbackup in a different location. Until then my unraid server is in a fireproof safe with a flotation device...

This is more aimed at backing up *other* computers to your Unraid server. https://www.urbackup.org Get a client from there and point at your Unraid server to back up to the server. Unraid's redundancy should mean you don't need to back it up to itself. (but I think some have managed to install the client onto Unraid itself...)

Images over 2TB - what do you do? I tried to backup a couple of 3TB drives to XFS formatted UNRAID system and got the error: Data on volume is too large for VHD files with 2.8054 TB. VHD files have a maximum size of 2040GB. Please use another image file format. As there is no other image format, is my only choice File Mode backups?

Err? All of that is probably why nothing worked. Ripping Win10 apart and downgrading it to a broken old protocol is not the best trick in the book. There seems to be a strange idea from some Linux hacks that dismantling security and putting back to the previous century's broken standards is a good idea. Win10 will work out of the box. Untouched. UNRAID is happy to talk Win10's language. Just set the share to private\secure\public as per the manual. Select the user to have read\write as needed. And all will be well. That's all I needed to do an I am no UNRAID expert. Using the IP Address is not a "back door". That is walking in the front door but without some daft human friendly name attached. The hack to get the username recognised doesn't seem too weird with WORKGROUPNAME attached. Think i have seen it before. Certainly see it with MACHINENAME\USERNAME paired. You are just telling Windows where to find that variant of a username when there is a clash. Or could be a knock on from all those other switches you hit at random. My main investigation would be retyping workgroup names \ check they are all in caps on all computers on the network, reboot all of them. These seems to imply a clash of usernames or slightly different WorkGroup Capitals. Are you logging into the Win10 box as a local user? Or with a Windows account? Both can mess with your SMB name in unexpected ways. Though you should still get the login prompt. I wonder if this is why being more specific on your username worked? I expect \UNRAIDSERVERNAME\USERNAME would also work in a similar way. Did you look into "Credential Manager" On win10? See what it had stored for that server? Including the capitals? Be aware that Win10 may well attempt to heal itself of those downgrading of the SMB security. I know I have turned on SMB 1.0 for old printer shares for people before, only for a bigger Win10 upgrade to come along and turn it back off again. I would strongly suggest undoing some of those old hacks and let Win10 and UNRAID talk clean SMB 2\3 or whatever they need. Trying to talk the old protocols will lead to confusions on both sides. Though ultimately - who cares how you got there. It works now.

Replaced my old post as I solved the puzzles. Instead a little bit of feedback from a noob to this plugin. First - I love the clear simplicity of this plugin. Thanks. Works great for a noob to UNRAID, and clearly lets me see which disks data is moving from. 1\ The GUI does not seem to like Vivaldi browser. A Chromium based browser. Had to swap to Firefox to get it to run correctly. Maybe my security settings a bit high? 2\ The pink text that pops up for a few seconds complaining about file rights is a little confusing. Not sure why Unbalance is worrying that folders were 770 instead of 777. Still worked fine for me. 3\ Why is there no security on this plugin? Anyone on my network could run this without permission checks. Means I'll need to keep it turned off when not in use.

I used midnight commander to move a single folder of many GBs. Well, I copied it to disk3 and then deleted it from disk1. Gaining those GB now seems to have the server happy again as "Common Problems" plugin is not complaining any more. Yay! Learning... now to get Unbalance to actually work. Currently it is stuck at the point of Select Disk. I go to GATHER, dig down a few folder to find a test folder to move, hit NEXT and select the target disk... but now it just sits there and won't light up the NEXT button. I'll give it a bit more of a poke, or will just go back to mc. okay - fixed that. Unbalance is fussy about web browsers. Only works in Firefox.

Now I have in interesting puzzle. That full drive is so full that UNRAID is mounting it in Read Only mode. Which gives me a Catch-22. I want to use the unbalance plugin to move data, but it can't move data that is stuck in read only mode. A read only mode caused by too much data that now needs me to move it... How do I tell UNRAID to set the drive to writable again so I can move data off of it to make space?

A couple of decades of working with SMB you learn the tricks. And where the real headaches lie. (Taming Windows 95 devices required use of a cattle prod) Haha. A Windows tool to help? Nah - just that pint of Guinness and a stop watch. My real suggestion is to look at what is on your network. What are the devices and who may be trying to boss into the election. Each time they start arguing you have the potential for two different answer coming back and your network getting split into two. You can also get the Master Browser falling asleep and needing to be woken up to reply with the list. I would sketch out a network on a pad of paper and look who is involved. Disconnect a few machines for a test. Does the network behave okay when it is just UNRAID, the Win10 machines and the router? Leave the Unbuntu disconnected. After a few hours of testing, and more Guinness, reintroduce some of the other machines in batches. Some devices just have different SMB implementations and cause confusions. Or could be causing arguments by having differing settings within their SMB setups. If you have been reading around, have you been doing registry tweaks to "fix" these issues? Often those tweaks are very out of date and can cause more problems than needed. My current home network is a mix of Win10, Unraid, Pi, TV boxes, HiFi, Printers all bickering over that same SMB network. I let UNRAID take over as Master Browser. No tweaks on any of the other machines and that Win10 box connected first time to Unraid shares over SMB. During the test, when you reintroduce machines, give them 30mins to settle. Crack another beer and walk away before testing. I am not saying the Unbuntu is "wrong", just it may be talking a slightly different flavour of disruption. This is one reason I think it is worth trying to get the UNRAID server to take over as master browser by hitting that tick. If one machine is fixed as master it stops the elections and means the Win10 boxes know who to talk to directly every time. (I am assume you have an always on UNRAID box) It could also be that your Master Browser is a PC that just falls asleep a lot. So when you try and browse it is needing to get a kick to wake it up causing that delay. Also I'll add something to @Frank1940 - check your cables. Any old cheapo Ethernet cables on the important machines? Try swapping them out. Only last month I had a cable fail after sitting there happy for 20 years... Or then there was that office where I found an ancient cheap 10BASE-T hub still in use complete with leaking capacitors inside... Sometimes it is the simple things

When I say "full" I mean the first 4TB drive is now jammed to within 20.5 KB of max, second drive is 10GB short of full. And I have Fix Common Problems nagging me. Also my OCD wants to see things tidy and levelled out a bit. Main excuse is it also gives me an excuse to mess around with unBalance as that looks quite a nifty tool. I find I learn this stuff the best by giving it all a good poke with a sharp stick. By looking into this it is letting me learn more about how neat the whole system is under the hood.

Thank you. Unbalance plugin looks good. Will dig more later once the new HDD is pre-cleared. Already poking at mc. Gives me flashbacks of XTree from DOS. Looks perfect for the task. Am I correct in my understanding that apps that run in Docker (i.e. urbackup) just see the unRAID drive as /mnt/user/ and it would be invisible to them if I manually move their data around between /mnt/disk1/ and /mnt/disk3/. (As long as stop that app when moving data).

Noob error. I have two 4TB drives in the box (and 12TB parity). Due to some bad maths my first backup I sent to it filled the drives both full. Eeek I now have a brand new 8TB drive to add to the UNRAID box. Is there a simple way of spreading out the load again? Manually moving 2TB from each of those 4TB drives onto the new 8TB drive? Or do I just need to get down with a command line and "mv" folders from \mnt\disk1\ to \mnt\disk3\ ? Or am I overthinking this? I know I have a few days of preclearing to do on the 8TB drive now, giving me time to plan. I also know I can just wipe the backup and start again, but that will take 3 days to repeat the backup. Thoughts? Hints?

You should be hassling that company too. Get them to update their outdated products. Remind them the days of XP are long gone. :D I abandoned them long ago due to that kinda attitude. Have you tried just typing the \\ip-address\ path in to their file explorer? See if it will jump to the right place? I'm kinda guessing they are just using very out of date libraries in their software if it doesn't properly handle a network UNC path.

I find Windows network is like Guinness - you have to wait for it to settle... 🍺 This is why I often use the IP address and bypass the whole naming scheme. I've always found the delays are just stuck in the naming lookups and you can go straight there with the IP Address. I like the fact that UNRAID notes that you can "cheat" Windows with user credentials as it Windows treats \\NAME as a different entity to \\IP-ADDRESS. As you note, SMB goes back to Windows for Workgroups and therefore predates the TCP\IP stack in Windows! 🎺

When restarting a Windows PC they should shout on the network "who is master server?". They check the OS level of that machine and may kick off an election if they are a newer \ higher OS version. That is why there is always a delay on start up. And it can get messy as it can take 20mins to sort out that new argument as all the machines may start bickering then. Not sure how this works with all the Win10 versions and SMB2. It was easier to make sense of when a network was Win XP Home\Pro and Win 7 Home\Pro. I am fairly certain having the UNRAID server set as master should keep things more stable. Having only just added UNRAID to my house I am yet to see the difference. (It certainly hasn't broken anything) It is also possible to edit the SMB.cfg file on a linux server to set a higher level so it kept as master. Certainly best to have the master on an always awake machine. (Been fighting these oddities for a few decades. LOL)

Does it work if you use \\IP-Address ? Quicker without the names? Guess by me is an issue with the Master Browser. Arguments on the network as to where to find stuff. The 30 second delay sounds like a slow name lookup, maybe kicking an election to sort out the Master to change and wake-up and reply. Things are not where the Windows machine expects to see them. Too many questions to ask. 1\ On the UNRAID server did you tick the "make master browser option"? Try doing the opposite setting. 2\ Is every computer on the same Workgroup - and be careful as WORKGROUP and Workgroup can sometimes be different. I always stick to ALL CAPS 3\ On the Windows machine what do you see in Network on Explorer? Do you see all the machines listed or are some missing? A network with a mix of OS will cause some possible confusions. Slightly different interpretations of SMB rules. There can be a 20 mins lag when changes occur to machine names \ power on for that to migrate around. You have a "master browser" who holds a list of where the machines are, and each other PC should ask the Master browser what is going on. A confused network can get two separate Master Browsers. Usually a guess is made to "elect" the "newest\biggest" OS as the Master to hold that list. Not sure how things behave with an UNRAID server saying "I am Master", It should make things run cleaner.

Users are always a pain to any nice network. What I do is make use of the Quick Access in the top left side of Windows Explorer. Pin their network folders in there. And for the real oldies pin a shortcut to the desktop. Use Explorer\Options to make it open to Quick Access (usually the default). That then deals with most people are looking for a X: drive I've avoided using Mapped Drive letters since the XP\DOS days. Pinning shortcuts around for the users is normally good enough for most in these GUI days.

So I guess what I have learnt today is to skip pre-read and post-read on new disks... the first of the 4TBs have just flipped over to the Post-Read stage 24 hours after turning the box on for the first time. Meaning the 12TB may be ready by Wednesday? 12TB has taken exactly 24hrs just to get through Pre-Read! I am hoping it will speed up a bit once the other two drives have completed their post-reading. And thanks for the formatting confirmations. Much appreciated. Edit: Little update for any other noobs reading. On a 3GB\s SATA2 system it took 3 days to complete the pre-clear. All three disc happily running at the same time, with the 12TB finishing up after 70.5 hours. Formatting was instant, and there was another 23hours wait after that for the parity to build. So after four days we have a happy server. Now onwards to sorting out urbackup...

Thanks for confirming that. So I can just buy BASIC today as this box maxes out at six drives and be happily using the same full SMB support as in the PRO. With the pre-clear, I assume now I have started I can't skip out of it without other consequences. I'm patient. One of the help files seemed to imply that the zeroing part of the actions are run when formatting anyway. I just hope I don't re-trigger that by a noob blunder. Haha. So, to check, some point later this week when the pre-clearing has finally completed, my steps are to Allocate the drives on the MAIN page - parity, disk1 and disk2. And then is it that same page I'll find an obvious FORMAT button? And let it format to XFS?

First - I am liking what I see with UNRAID. Initially aiming to use this as an urbackup server. Setup going well, even if some of the documentation is a little confusing \ out of date. Two different questions: 1\ I want to make use of SMB user rights fully. Different users will need different access to the directories on the server. Do I have to buy a PRO licence for this? Or will BASIC be okay for full SMB use. I see a confusing contradiction in the documentation and my search abilities have failed me. 2\ Pre-clearing drives. Wow! is it really supposed to be this slow? I have an old ProLiant N40L server, 8GB RAM, drives running at 3Mbps, and I have started with a 12TB (parity) and two 4TB drives in there. I installed the handy Pre-clearing addon so I could get at this from the web GUI. I set going on all three drives... ouch. 22 hours later and the 12TB is still 94% through pre-reading, and the 4TBs are 94% through zeroing. At that rate it implies to me that it will take me three days to get the 12TB zeroed! Please tell me that the next steps aren't as long! Seeing "step 2 of 5" is a little scary. BTW - these are brand new, never used, WD Reds. More disks will be thrown in at a later time as this is just a start.