Unraid Future Feature Desires Poll

[enJOyIT]

10 minutes ago, ChatNoir said:

If you revert a snapshot, is it not equivalent to a backup ?

In my opinion a backup is the possibility to store the backup file at another storage(s) (other NAS, Array etc). And store multiple versions (e.g. every day/week/etc, keep the last 5/x versions of that backup).

 

As I mentioned proxmox did a very good job at this point. So why don't use it as an example.

Edited July 23 by enJOyIT

[SimonF]

19 minutes ago, ChatNoir said:

If you revert a snapshot, is it not equivalent to a backup ?

No as it will not longer be available to use again once reverted.

[SimonF]

6 hours ago, enJOyIT said:

I thought it was a point in 7.0

I dont think backups where announced to be part of 7.

 

PBS is an enterprise class solution so suspect lots of money and time was tken to develop.

[enJOyIT]

8 hours ago, SimonF said:

I dont think backups where announced to be part of 7.

 

PBS is an enterprise class solution so suspect lots of money and time was tken to develop.

I don't talk about the PBS. I mean the integrated backup/restore mechanism in proxmox hypervisor...

Edited July 24 by enJOyIT

[Espressomatic]

On 7/23/2024 at 11:41 AM, ChatNoir said:

If you revert a snapshot, is it not equivalent to a backup ?

 

No. Snapshots are not backups. Kind of in the same way that a mistress is not a backup of a wife.

That said, I plan to use 3rd party backup software which will hopefully also take care of the VMs. (instead of making manual copies like I have been now)

 

 

Edited July 24 by Espressomatic

[Espressomatic]

On 7/19/2024 at 9:04 PM, WizP said:

Would love to be able to have VMs/dockers start without the array needing to start. 

This would help when using VMs for routing software. (I know many people don't like it but works perfectly for homelabs and my use case) 

This is still not possible in 7 betas, so I'm not sure what the other replies were about.

 

"Stopping the Array" (Array Operations -> Stop) means stopping all Unraid services and shares, including VM and Docker, not just the pool of disks known as the "array."  It makes no difference where data is stored for VM or Docker, because Stopping brings those services down completely. That is, when it works. Recently this has been hit or miss in testing - stopping will typically hang up forever, unable to bring the shares down after taking down the other services.

 

I hope that the entire list of buttons/functions under Array Operations goes away, and that specific functions are moved to more relevant pages/tabs ion the UI. Like Shut down, Reboot and Sleep going into a power menu/button (and they're already on the main dashboard). Spin up/down moved to the disk sections/tabs along with Move.

 

I've split my services and functions of my old server onto multiple different machines, so I never have to use the start/stop buttons. My apps/services system has no array-type pool and is all solid-state ZFS. My NAS system has no docker nor VM and is mostly spinners in ZFS. Routing/DHCP, Firewall and DNS has its own mini system and as of tomorrow Home Automation on its own system too.

 

 

Edited July 24 by Espressomatic

[Faceman]

On 7/22/2024 at 8:34 AM, Doppler said:

More Mover features would be amazing. Specifically on a share by share use case. Say I want to have a share for all my freshly downloaded media to only move over after a week, while I only want my CCTV footage to move over once a day, my "working" share to only move data from cache to array every month. That would be invaluable for me. It would optimize cache usage IMO. Instead of having to manually setup cache > array and run Mover overnight on my working array and keeping CCTV/New Downloads for a week. You can set data intensive loads to go quicker, smaller loads to go later. 

Are you already using the popular Mover Tuning Plugin to do this?

 

I use mover tuning with two separate caches and have difference settings per share, mover runs hourly in my case but only moves files that are 2 weeks old or if the pool gets full.

 

the only thing it cant do at the moment is move data in chunks when full, it can only trigger a full dump move at a certain threshold. I'd like to have more of a constant rotation, keeping it filled to a certain set percentage automatically and moving to the array as needed.

 

some features I'd like to see on top of that being integrated would be multiple tiers, for fast cache (NVME pool for example), storage (e.g. ZFS HDD array) then an archive of spun down HDDs in an unraid array.

 

And some kind of read caching, even a simple folder preload script, for example, you read a file from the HDD array and it automatically copies the rest of that folder up a tier to the SSDs, up to a certain threshold and automatically clears it if not used within a set time frame for example.

 

 

Edited July 25 by Faceman

[Espressomatic]

On 7/23/2024 at 10:28 AM, enJOyIT said:

Look at proxmox and you know what a nearly perfect backup/restore solution is. I thought it was a point in 7.0

 

If that's true, I don't know how. Is it made 100% by different developers than work on PVE? Because PVE is, without reservation, one of the worst server-based applications I have ever had the displeasure of using. It's 100% garbage and I have to restrain myself from getting much more descriptive. If I never have cause to install it nor access its UI ever again, I'll be very content.

 

I had given up on anything PVE except for one single small system that was to be dedicated to nothing by Home Assistant VM and a few LXCs for platforms that tie directly to Home Assistant and/or Homekit, like Scrypted, MQTT, etc.  Had everything installed (using scripts because that's the only way to use PVE) and the whole system crashed after some time while I was working in Scrypted.

 

Unraid has never crashed and brought down my entire server since I first set it up in 2018.  So, that one simple system with only a single NVME drive in it, is also going Unraid, just to host the one VM.

 

As for backup...  I was planning to set up Borg, which also has a number of clients for various (other) platforms.

 

 

 

Edited July 25 by Espressomatic

[enJOyIT]

Sorry that it is a bit too much rant from you.

 

Proxmox is imho the easiest enterprise class hypervisor at the market. 

 

And trust me the integrated backup in pve is far more easier than everything else in unraid and it's plugins "crap". And it is stable as f***. Backup while VM is running? No problem... Even restore is so easy. And it's build in! 

 

So don't talk about things you don't know.

Edited July 26 by enJOyIT

[Rysz]

Can you guys chill out a bit please, after all everyone's entitled to their own opinion in this topic.

Nobody's forcing anyone to use our plugin "crap" either, so feel free to use whatever works best for you.

[SpencerJ]

3 hours ago, enJOyIT said:

plugins "crap". And it is stable as f***. Backup while VM is running? No problem... Even restore is so easy. And it's build in! 

 

So don't talk about things you don't know.

 

Agreed. Please tone it down. Confused? Review our community guidelines.

[Espressomatic]

3 hours ago, enJOyIT said:

So don't talk about things you don't know.

30 year software designer. 

 

As I understand, we're here in this forum because Unraid is well suited to whatever use case we're putting it through.

If you want to use Proxmox, knock yourself out. I would never bring it up as any kind of (positive) example with respect to Unraid's feature set nor future plans. And as far as Unraid backup solutions, there are some compelling alternatives out there already, a number of which have been mentioned.

Edited July 26 by Espressomatic

[enJOyIT]

Sorry for the emotional post.

 

But naming proxmox as "worst server-based applications" was triggering me.

[Espressomatic]

The technologies and software under it (which they didn't write) are great. Just too bad about the front of house.

 

Steak in the hands of some people just ends up hamburger. Another case in point is OPNSense. The software it's based on, pfSense is so much better (and usable) in every way. For a few weeks I thought I was living on backwards planet Bizzarro because the opn fanbase is so rabid and in love with it. One might say almost cult-like.

Edited July 26 by Espressomatic

[salora]

Proxmox debate set asside it is not the job of the virtualisation software to handle backup. I am sys admin and at work i don’t ask vmware to handle my backups i have veeam for that. So at home i don’t ask unraid to backup my vm’s.

the free veeam agent can handle that for example. That being said snapshots are a great feature

[enJOyIT]

Are you backup the .img files with veeam?! And where is veeam running? In an VM? What is when your host has a problem running that VM? How do you restore backups?

[JonathanM]

2 minutes ago, enJOyIT said:

Are you backup the .img files with veeam?! And where is veeam running? In an VM? What is when your host has a problem running that VM? How do you restore backups?

I use UrBackup, I'm assuming the exact same way they use veeam. The VM is running the backup agent, so the backup agent has full knowledge and control of the file system. You get a good backup while the VM is up and running, because the backup agent can do shadow copies, cache flushing and other manipulations to ensure backed up files are valid.

 

When the VM is borked, you change the boot iso to the backup software's recovery agent, and retrieve the backup image, overwriting the borked files.

 

If you back up .img files, you have several options, only one of which is ideal. You can shut down the VM so the backup is in a fully consistent shut down state, you can back it up while the VM is running unaware, which means the backup is functionally equivalent to what you would get by pulling the power and crashing the VM, or you can set up the kvm guest agent and use the proper function calls to get the OS to commit changes to the image.

 

Until all the pieces are in place for the last option in Unraid, the only real solution is running the backup agent inside the guest.

 

This solution works now, doesn't need anything extra, and is very reliable, I've moved VM's to a new host using UrBackup.

 

I'm not saying native qemu backup in Unraid isn't a good ask, and it's reasonable to request it, just that it's not the only option, there are many good backup methods for VM's that work now.

[enJOyIT]

Yeah I know UrBackup and other solutions, but I don't want any "bloatware" in my guest machines to do a backup. Nor I want to run a VM which runs a backup solution and I am in trouble when my VM-system is broken or has other problems.

 

My last statement for this topic... A simple built-in backup solution is a must have for an hypervisor. I'm just asking limetech to look at the solution in Proxmox (not the Proxmox Backup Server! I mean the built-in solution of the hypervisor). It is a really simple approach and in my opinion that shouldn't be a big task to do it in a simlar way.

[Doppler]

On 7/24/2024 at 8:49 PM, Faceman said:

Are you already using the popular Mover Tuning Plugin to do this?

 

I use mover tuning with two separate caches and have difference settings per share, mover runs hourly in my case but only moves files that are 2 weeks old or if the pool gets full.

 

the only thing it cant do at the moment is move data in chunks when full, it can only trigger a full dump move at a certain threshold. I'd like to have more of a constant rotation, keeping it filled to a certain set percentage automatically and moving to the array as needed.

 

some features I'd like to see on top of that being integrated would be multiple tiers, for fast cache (NVME pool for example), storage (e.g. ZFS HDD array) then an archive of spun down HDDs in an unraid array.

 

And some kind of read caching, even a simple folder preload script, for example, you read a file from the HDD array and it automatically copies the rest of that folder up a tier to the SSDs, up to a certain threshold and automatically clears it if not used within a set time frame for example.

 

 

Not using it at the moment, it doesn't really fit with what I need. I don't want it moving the files by age, just on a daily/weekly/monthly routine. If I moved them by age I would never achieve my 80 percent standby rate for my drives. I prefer moving it in bulk after the files "lifespan" has passed. (How long I typically work/need it immediately) I add files every hour or so, and as far as I know it would just keep moving them over and never let my drives spin down. Maybe it can be set up differently but I do not know how. That's why a share by share use-case is far more important. Integrated functionality is always good. Maybe a "Pool" Generalization/Default setting that can be overridden on random shares. That would make it nice for different cache pools that take in different types of data. Which would affect all the shares using said pool. Just overkill but a more convenient initial setup. That was a whole word salad of a post, sorry, am a bit exhausted.  

[enJOyIT]

Regarding this thread:

An option to disable CPU pinning would be nice. 

[Veah]

ARM support

[bmartino1]

2 hours ago, Veah said:

ARM support

ARM support is in the kernel already it the x86 package executables and packages for arm that is the problem.

Not all x86_64 packages exist for arm code. This would asl double the space on the USB to carry an arm version over x86_64 version.

I could see a separate unraid version for arm fork using new arm code, but not arm support on current unraid generation.

[Veah]

1 hour ago, bmartino1 said:

I could see a separate unraid version for arm fork using new arm code, but not arm support on current unraid generation.

That would be great!

[keithcroshaw]

A decent mobile UI, at least for simple docker tasks? When stuff goes south and you're in couch mode pinching to zoom, opening header tabs in new browser tabs because clicking them doesn't navigate, it gets old. I usually just groan and get the laptop, most times it's in my basement office though. 

[JohnHedge]

User defined Docker network support in the UI