Hello

 

Apologies if this has been answered already.

 

Is it possible to prevent the UNRAID GUI from wakening a HDD from sleep when assigned using 'unassigned devices' plugin?

 

I have 2 4TB HDDs (sde and sdf) that I set to sleep after 2 minutes of no activity using hdparm -S $device

However, after they sleep and I refresh the UNRAID main page that shows the disks, both drives spin up. Is this normal?

Is there anything I can do to stop this from happening?

 

I tested this using hdparm -Y /dev/sde forcing the drive to sleep and then refreshing the UNRAID gui. This wakes up the drive also.

 

Anyone using a script to put their drives to sleep and works OK, please share it so I can see the script config. Thank you.

 

Read the second post about best practices.

Is AFP share support on the roadmap for Unassigned Devices?

 

No.

 

What are the best methods/practices for using Time Machine via AFP on a non-array disk? I prefer to avoid using Time Machine via SMB.

 

Has anyone set up Time Machine within an array share and symlinked it to a non-array mounted drive controlled by Unassigned Devices?

 

Why do you want to do this?  The UD mounted device is not protected by parity.

 

What are the best methods/practices for using Time Machine via AFP on a non-array disk? I prefer to avoid using Time Machine via SMB.

 

Has anyone set up Time Machine within an array share and symlinked it to a non-array mounted drive controlled by Unassigned Devices?

 

Why do you want to do this?  The UD mounted device is not protected by parity.

 

Time Machine for my usage case is only backup data; which is only important in the event that my workstation HD dies or I lose files. My backup data is easily replaceable by creating a new backup if necessary. I don't want to waste valuable array storage and would rather keep a large dedicated Time Machine disk outside of the array. 

 

What are the best methods/practices for using Time Machine via AFP on a non-array disk? I prefer to avoid using Time Machine via SMB.

 

Has anyone set up Time Machine within an array share and symlinked it to a non-array mounted drive controlled by Unassigned Devices?

 

Why do you want to do this?  The UD mounted device is not protected by parity.

 

Time Machine for my usage case is only backup data; which is only important in the event that my workstation HD dies or I lose files. My backup data is easily replaceable by creating a new backup if necessary. I don't want to waste valuable array storage and would rather keep a large dedicated Time Machine disk outside of the array.

I do not understand the logic. If you have the disk in the server anyway, it doesn't matter if the disk is in the array or in UD. You still have the same amount of space available.

If you worry about the time machine share using more than one disk, just make the share only include the one disk.

 

What are the best methods/practices for using Time Machine via AFP on a non-array disk? I prefer to avoid using Time Machine via SMB.

 

Has anyone set up Time Machine within an array share and symlinked it to a non-array mounted drive controlled by Unassigned Devices?

 

Why do you want to do this?  The UD mounted device is not protected by parity.

 

Time Machine for my usage case is only backup data; which is only important in the event that my workstation HD dies or I lose files. My backup data is easily replaceable by creating a new backup if necessary. I don't want to waste valuable array storage and would rather keep a large dedicated Time Machine disk outside of the array.

I do not understand the logic. If you have the disk in the server anyway, it doesn't matter if the disk is in the array or in UD. You still have the same amount of space available.

If you worry about the time machine share using more than one disk, just make the what only include the one disk.

 

My array slots/devices are maxed out (can not add any more disks to the array) with the purpose of storing media and other data. Additionally, there is no room to add other drives inside the server chassis. Unassigned Devices makes it easy to add additional external disks via USB. I want to dedicate my large external disk as Time Machine data.

 

I'm feeling mentally switch-tracked by having to defend my particular use case. My array is fully allocated. I have an external non-array USB drive mounted using Unassigned Devices that I intend on using for Time Machine backups over the network. I am looking for a solution not involving SMB. Since UD doesn't support AFP, I suppose this plugin won't help. It's a very useful plugin otherwise.

 

I'll attempt to symlink within an AFP share (from an array device) to a non-array mounted device. This wouldn't inherently depend on the UD plugin, so it is outside the scope of this thread.

 

What are the best methods/practices for using Time Machine via AFP on a non-array disk? I prefer to avoid using Time Machine via SMB.

 

Has anyone set up Time Machine within an array share and symlinked it to a non-array mounted drive controlled by Unassigned Devices?

 

Why do you want to do this?  The UD mounted device is not protected by parity.

 

Time Machine for my usage case is only backup data; which is only important in the event that my workstation HD dies or I lose files. My backup data is easily replaceable by creating a new backup if necessary. I don't want to waste valuable array storage and would rather keep a large dedicated Time Machine disk outside of the array.

I do not understand the logic. If you have the disk in the server anyway, it doesn't matter if the disk is in the array or in UD. You still have the same amount of space available.

If you worry about the time machine share using more than one disk, just make the what only include the one disk.

 

My array slots/devices are maxed out (can not add any more disks to the array) with the purpose of storing media and other data. Additionally, there is no room to add other drives inside the server chassis. Unassigned Devices makes it easy to add additional external disks via USB. I want to dedicate my large external disk as Time Machine data.

 

I'm feeling mentally switch-tracked by having to defend my particular use case. My array is fully allocated. I have an external non-array USB drive mounted using Unassigned Devices that I intend on using for Time Machine backups over the network. I am looking for a solution not involving SMB. Since UD doesn't support AFP, I suppose this plugin won't help. It's a very useful plugin otherwise.

 

I'll attempt to symlink within an AFP share (from an array device) to a non-array mounted device. This wouldn't inherently depend on the UD plugin, so it is outside the scope of this thread.

 

It's hard for me to know that your UD disk is a USB disk 

You can still add the disk to your array though, as I think support for USB disks was added in 6.2. Then you can make it an AFP share for your Tardis.

 

Using a network share (even Apple's own Time Capsule) as a Time Machine destination is never as reliable as directly connecting the disk to the Mac. The difference is that in order to implement Time Machine across a network a sparse bundle disk image has to be created on the destination, whereas a directly connected hard disk can be formatted appropriately. Sparse bundle disk images consist of many many small files and they have proven to be rather fragile in practice. What I do is directly connect an external USB-connected disk to each of my Macs and let Time Machine write to it directly. It's simple and reliable.

Hi Dan,

 

Is it possible with UD as it stands to export a mount as NFS but with specific parameters, much like can be done with regular shares when exporting as NFS?  I'm guessing not, but figured I'd ask.  Admittedly it's a bit of an edge-case scenario, but would be very helpful to me at least if it was possible to tweak the NFS export parameters, rather than standard Public with read/write access.

 

I guess also that would mean having to change NFS export from a global setting to per-mount, as I notice that /etc/exports- get's populated with all mounted devices once NFS export is set to enabled.

 

I was going to take a look under the hood at your code and see if I could find where /etc/exports- gets written.  Before using UD, I used to have a manually created exports- at /boot/config which got auto-added to /etc/export when exportfs was called.  Is there something similar perhaps, whereby you store a copy of exports- or do you generate during plugin run only?

 

I could go out an manually change /etc/exports- myself, but that's messy for all kinds of reasons. 

 

Would appreciate your thoughts.

 

 

Regards,

overbyrn

One of my unassigned devices didn't unmount on array stop, had to start the array again (disk wasn't exported) and stop a 2nd time to unmount, there were no files open on that disk at the time of the 1st stop.

tower7-diagnostics-20161201-1618.zip

One of my unassigned devices didn't unmount on array stop, had to start the array again (disk wasn't exported) and stop a 2nd time to unmount, there were no files open on that disk at the time of the 1st stop.

 

Look at the UD log and see what prevented the unmount.

Look at the UD log and see what prevented the unmount.

 

There's nothing there.

Hi Dan,

 

Is it possible with UD as it stands to export a mount as NFS but with specific parameters, much like can be done with regular shares when exporting as NFS?  I'm guessing not, but figured I'd ask.  Admittedly it's a bit of an edge-case scenario, but would be very helpful to me at least if it was possible to tweak the NFS export parameters, rather than standard Public with read/write access.

 

I guess also that would mean having to change NFS export from a global setting to per-mount, as I notice that /etc/exports- get's populated with all mounted devices once NFS export is set to enabled.

 

I was going to take a look under the hood at your code and see if I could find where /etc/exports- gets written.  Before using UD, I used to have a manually created exports- at /boot/config which got auto-added to /etc/export when exportfs was called.  Is there something similar perhaps, whereby you store a copy of exports- or do you generate during plugin run only?

 

I could go out an manually change /etc/exports- myself, but that's messy for all kinds of reasons. 

 

Would appreciate your thoughts.

 

 

Regards,

overbyrn

 

All UD devices that are marked as shared are shared with NFS in the same way - public read/write access.  The /etc/exports file is modified for the UD device whenever that device is mounted.

 

What is it that you want to be able to do?

All UD devices that are marked as shared are shared with NFS in the same way - public read/write access.  The /etc/exports file is modified for the UD device whenever that device is mounted.

 

What is it that you want to be able to do?

 

I'd like to be able to set NFS to private on a per UD device basis.

 

eg.

All UD devices that are marked as shared are shared with NFS in the same way - public read/write access.  The /etc/exports file is modified for the UD device whenever that device is mounted.

 

What is it that you want to be able to do?

 

I'd like to be able to set NFS to private on a per UD device basis.

 

eg.

 

The SMB share permissions are set on a global basis, not by UD device.  I can do the same thing for NFS shares.  UD was not designed to allow per device share permissions.

Hi all,

 

I'm new to unRAID (migrating from an ESXi setup) and I've having some issues getting my VMFS file system disks mounted/visible on unRAID. With the default unRAID setup, the file system on my disks was just coming up as "unknown". Now that I've added the "unassigned devices" plugin, the disks are visible with the proper file system detected (VMFS).

 

I would typically just format the disks and start with a clean slate but all of my media data is stored on these disks as well as ESXi VMs. My goal is to get these disks mounted and visible on an unRAID Ubuntu or Win 10 VM so that I can extract the data and move it to newly provisioned array disks.

 

When I try to mount these disks in unassigned devices, the mounting process cycles for a few seconds but the mount button never changes to unmount... so I take it the mount is failing at some point of this process. I'm guessing that all disks should be mountable regardless of the file system?

 

If it matters, these disks are attached via SATA to an LSI 9240-8i that's been cross flashed to an 9211-8i in IT mode.

 

Thanks for any help you guys can provide!

 

Chris

I'm guessing that all disks should be mountable regardless of the file system?

 

That doesn't necessarily follow. For a file system to be mountable it has to be supported. This plugin loads support for some file systems but, it seems, not VMFS.

 

I would typically just format the disks and start with a clean slate but all of my media data is stored on these disks as well as ESXi VMs. My goal is to get these disks mounted and visible on an unRAID Ubuntu or Win 10 VM so that I can extract the data and move it to newly provisioned array disks.

You could try creating the appropriate VM and then passing the physical disk through to that VM.  Done that way the VM would handle the mounting so it does not need to be mounted at the unRAID level.

Thanks guys, I thought in order to pass through a disk to a VM it needs to be in the array as an assigned disk? When I look at available disks on my Ubuntu VM I only see the local disk that was provisioned for it.

 

If I open up terminal and take a look through fdisk, I only see that disk as well. Mount points also do not reflect any of the unassigned disks I'm looking to mount.

 

Perhaps I'm missing a simple step but how do you go about passing through one of these VMFS disks right to a VM without mounting in at the unRAID level? I suspect editing the VM config file can accomplish this?

... I thought in order to pass through a disk to a VM it needs to be in the array as an assigned disk?...

You have this exactly backwards. If you pass the physical disk to a VM it is unavailable to unRAID so cannot be an assigned disk.

To pass through a physical disk to a VM you use its device ID (I.e. Something along the lines of /dev/sd? Where ? Identifies the disk in question.  You can also (which is more robust as the /dev/sd? Type identifiers can change between boots) pass it through using the /dev/disk/by-uuid/? type names.

That's the strange part, I'm fairly certain the physical disks (presented as sda, sdb etc in linux) were not appearing the dev folder. Does it matter how these disks are connected? They are directly connected to an 9240-8i which has been crossed flashed to IT mode 9211-8i.

 

I was expecting to just see these disks in the Ubuntu VM's /dev folder but no such luck. That's why I attempted to mount these VMFS disks using the "unassigned devices" plugin in hopes I could mount them in unRAID and view the contents that way.

 

Thanks again

I was expecting to just see these disks in the Ubuntu VM's /dev folder but no such luck.

If there are no unraid disks on the controller, you could pass the entire controller through to the VM, and it would work the way you are expecting. If you can't pass the whole controller, you will need to pass through the individual disks like itimpi said.

I'm definitely understanding what you guys are suggesting it's just that the hard disk passthrough component doesn't appear to function as expected.

 

To simplify troubleshooting this, what's the best way to passthrough the whole controller to a VM? When creating a VM there are only a few devices that appear in the passthrough device option list (video card, USB controller etc.). Do I need to edit the VM config file and specify the SATA controller for passthrough that way?

I have a question if Unassigned Devices plugin can help me.

I have two hard drives: 3TB HDD (with data on it) and 4TB HDD (brand new, empty disk). When creating my array for the very first time, I picked from the dropdown list both of them. Next I noticed I have to format them in order to use them in array, so it was a no-go... so I then went to Tools and started brand new config.

I was able to create new array with just 4TB, formatted it and was able to work with it. My question now, how can I access my 3TB disk with data so I can copy everything to 4TB disk and then format 3TB disk and copy some of it back.

 

I just removed my 3TB disk and placed it to my other PC, but unfortunately in Win10 I can't see it in My Computer (I can see it in a place where you manage disks, but can't do anything). Will I be able to still access my data somehow? I didn't format 3TB disk in unRAID.

 

After installing the Unassigned devices plugin I can see my 3TB disk on the list, but there is no 'Mount' option beside it :<

I have two hard drives: 3TB HDD (with data on it) ...

 

How is it formatted?