unRAID Server Release 5.0-beta1 Available

[limetech]

Download

 

Please note... this is a Beta.  Even though there have been no significant changes in the unRAID driver, I recommend running -beta1 on a "test" server.

 

Also note: there is a more involved update procedure... see below.

 

Please use the new message board section specifically for this release.

 

 

unRAID Server 5.0-beta1 Release Notes
=====================================

New Features:
- New extensible webGui architecure.  Pages are rendered with PHP, but other methods are possible.
- New security model.
- Security is now available in 'Basic' verion.  The feature differences between the different license key types are as follows:
  Basic - support 3 array drives, no cache, no Active Directory support
  Plus  - support 6 array drives plus cache drive, no Active Directory support
  Pro   - support 21 array drives plus cache drive, Active Directory support
  
Other:
- Beta-1 release does not have all functionality yet:
  No NFS.
  No AFP.
  No FTP.
  No Active Directory (actually may work, not tested).
  No "official" plugin scheme designed yet (this will be done with help from Community).
  No event hooks for plugins (again, will be implemented with input from Community).


Changes from 4.5.6 to 5.0-beta1
-------------------------------

- Way too numerous to list. Only minor changes in the unRAID driver itself, so data transfer integrity should be safe.


Upgrade Instructions
====================
Please read carefully.

Because of the new share security model, some additional steps are necessary to upgrade from version 4.5.6:

1. Copy the 'bzimage' and 'bzroot' files from the zip file to your flash device.

2. If you are upgrading 'on-line', Stop the array and then delete these two files from the flash:
    config/passwd
    config/smbpasswd
    
(Sorry, you will need to re-create all your users after booting 5.0-beta1).

3. Reboot the server.  When server has booted, after examining snazzy new webGui, navigate to the 'Utils' tab and select "New Permissions".  This is a one-time operation to fix the file and directory permissions.

 

[NAS]

First "this is awesome well done" post.

 

I expect many more.

 

Kudos

[limetech]

Wow, you're fast, did you already try it?

[MvL]

Thank you Limetech!

 

We have testing to-do 

[speeding_ant]

Man that's awesome! Well done. I'd love to try it out, but don't have a test server to test it on. Am looking forward to AFP immensely.

[limetech]

Am looking forward to AFP immensely.

 

AFP support is the main reason for the change in security model.

[speeding_ant]

So will there be an API for us? I'm keen to create some nice PHP add-ons 

[Joe L.]

Tried this on my test server.   First screen I'm presented with looks like this (seems difficult to read  ):

 

Edit: Solved...  Do NOT install php on your own.  (I had the "php" package installed via unMENU.  It was interpreting the pages instead of the lime-tech supplied code in emhttp.)

To fix, go to the package-manager page in unMENU, disable the re-install on reboot for the "php" package

Then, go to the Array Administration page in unMENU and stop the array

Last, log in on the command line and type:

reboot

[squirrellydw]

I have to ask, will 5.0 add a 2nd parity drive?

[neilt0]

I have to ask, will 5.0 and a 2nd parity drive?

 

Not yet. It's right in the first sentence of the first post:

 

"there have been no significant changes in the unRAID driver"

[jupilerman]

Looking good.

 

No refresh button ==> hit the browser refresh does the job.

Refresh is faster then 4.5XX

 

[Guest wolfdv]

sweet, will try when i get home

[bubbaQ]

I created a new array from scratch with a spare drive I had lying around, that was partitioned into 3 ex2 partitions and a swap, and had other data on it.

 

When I created an array with it, it did not clear it first, but only requested a format, and then the array started right away.  Is that expected behavior?  Will it only be adding a drive to a protected array that will invoke clearing?

[Joe L.]

I created a new array from scratch with a spare drive I had lying around, that was partitioned into 3 ex2 partitions and a swap, and had other data on it.

 

When I created an array with it, it did not clear it first, but only requested a format, and then the array started right away.  Is that expected behavior?  Will it only be adding a drive to a protected array that will invoke clearing?

I don't think unRAID has ever cleared a disk if parity is not yet assigned, on any version.

[bubbaQ]

Makes sense... I just never noticed before because I always preclear.

[Joe L.]

A small issue, but not with unRAID 5.0.   So far, I'm impressed.

 

There will be issues if you had existing "php" add-on applications.  unRAID 5.0-beta1 breaks the smarthistory.php program.

 

If you are using unMENU you'll want to use its built-in feature to update itself.  the new 5.0-beta version of unRAID changed how /proc/mdcmd behaves when a parity check is under way.   This causes unMENU to incorrectly report that a parity check is in progress when it is not.  (The variable used to be equal to "" if a parity check is not in progress, in 5.0-beta1 it is now set to "0"  The update to unMENU now checks for either condition.

 

Joe L.

[BRiT]

Nice to see it posted, will be trying it out later today.

[BRiT]

Even though my array have not grown beyond the limits of a basic setup yet, I wanted to show support for the development efforts of this milestone by finally placing my order for a Pro 2-Pack. I know this is a product I will still to use when my needs grow, so I look forward to continued product developments.

 

 

[internetfriend]

I have to ask, will 5.0 and a 2nd parity drive?

 

Not yet. It's right in the first sentence of the first post:

 

"there have been no significant changes in the unRAID driver"

 

I know its pretty much asking for a complete overhaul of the code, but this would definitely be something I'd look forward to also as a feature down the road.

[Nyago123]

+1 for second parity.  It's come up before in multiple places, but in addition to the UI improvements, it's what I'd be looking for most.

 

That also says that for the most part, I'm pretty happy with unRAID in general. 

[speeding_ant]

Tried this on my test server.   First screen I'm presented with looks like this (seems difficult to read  ):

 

Edit: Solved...  Do NOT install php on your own.  (I had the "php" package installed via unMENU.  It was interpreting the pages instead of the lime-tech supplied code in emhttp.)

To fix, go to the package-manager page in unMENU, disable the re-install on reboot for the "php" package

Then, go to the Array Administration page in unMENU and stop the array

Last, log in on the command line and type:

reboot

 

Could have something to do with the fact that <? isn't seen as a beginning tag on newer installations of PHP?

[bubbaQ]

The standard php opening is:

 

<?php

 

The form using by unRAID 5.0 is considered the "short" form of:

 

<?

 

You can control this with the short_open_tag directive in php.ini

 

You used to be able to override it like this:

 

<?php ini_set("short_open_tag", 0); ?> 

 

at the top of your code, and that would allow the short open from there on, but tat was removed in php 5.x.

[BRiT]

Thank's for pointing that out bubbaQ.

 

For those wishing to experiment with this on a full Slackware distro, here's a very rough outline of the steps I took to upgrade mine from unRAID 4.5.6 to 5.0 Beta 1. When I refer to relative paths, it is rooted at 'unraid-5.0beta1' directory.

 

Create new Linux Kernel

 

1. cp usr/src/linux/drivers/md/* /usr/src/linux/drivers/md/

- updated md.c file

 

2. rebuild kernel

- rebuild kernel with updated md driver

 

3. reboot into new kernel

 

Update from unRAID 4.5.6 to unRAID 5.0

 

1. cp usr/local/sbin/emhttp usr/local/sbin/shfs /usr/local/sbin

- updated emhttp

- updated shfs

 

2. cp -R usr/local/webGui /usr/local/webGui

- updated web pages

 

3. mkdir -p /var/local/emhttp /var/local/webGui /var/lib/php

- updated working directory emhttp

- updated working directory webGui

- updated working directory php

 

4. cp etc/rc.d/rc.samba /etc/rc.d/rc.samba

- updated samba script

 

5. cp etc/samba/smb.conf /etc/samba/smb.conf

- updated samba configuration, remove {passdb backend = smbpasswd}

 

Post install Tweaks

 

1. edit php.ini [/etc/httpd/php.ini] to allow '<?' to work, {short_open_tag = On}

 

2. edit go script to invoke emhttp redirecting stdout and stderr to a log file, since it was chatty to the console window.

/usr/local/sbin/emhttp -p 88 >/var/log/unraid/emhttp 2>&1 &

 

3. edit emhttp for module remove command, change 'modprobe -rw' to 'rmmod -w', since newer versions of the module-init-tools from version 3.10 onward no longer provide for the modprobe '-w' or '--wait' options, but still offer it on rmmod. *

 

* NOTE: Editing emhttp is not for the faint of heart, should be done with a binary editor, and not needed upon fresh reboots, but may be needed if you stop/start the array more than once.

 

 

[dlmh]

Am looking forward to AFP immensely.

 

AFP support is the main reason for the change in security model.

 

This is awesome, Tom. Could you elaborate a bit on this? Will it still be a problem to have user shares over AFP?

[pastrychef]

Am looking forward to AFP immensely.

 

AFP support is the main reason for the change in security model.

 

Thank you very much for this!!  I am extremely excited about the final!