unRAID Server Release 6.0-beta14b-x86_64 Available


limetech

Recommended Posts

I'm glad things are progressing jonp - I'm just about to rebuild my home server, and it's a toss-up between unRAID and OpenMediaVault.  OMV was winning the flip due to a lack of updates in unRAID.

 

I do like unRAID, it seems much happier to spin down it's drives compared to OMV, which is a massive bonus to me.

 

How are the Haswell throttling issues progressing?  The new box is a Haswell Xeon, and I'd like to see it idling at 800MHz as it should. :)

 

No new updates on that.  It's a low-priority fix for us at this point and will not hold up the release of 6.0 (it doesn't cause system crashing and other processor models remain unaffected by this).  I'm fairly confident this will get resolved in due time, but just can't be certain on when.

 

I don't believe it is just haswell xeons as my CPU is currently:

 

CPU utilization	
0%
CPU speed	core 1 / 2	3163 MHz	3126 MHz
core 3 / 4	2999 MHz	3093 MHz

 

it is an Intel® Xeon® CPU E3-1220 V2 @ 3.10GHz

 

Xeon is still considered part of the Haswell chipset.  You have a haswell.

 

NO !!    The E3-12xx v3  (NOTE the THREE in v3) CPU's are Haswell architecture CPUs.    The v2's are Ivy Bridge  :)

Link to comment
  • Replies 476
  • Created
  • Last Reply

Top Posters In This Topic

... should have read the rest of the thread before posting -- I guess it's now VERY clear to anyone reading this thread which Xeons are Haswell  8)

 

My bad.  So I guess it affects Ivy as well then?  I thought we tested an Ivy system here and didn't see the issue, but will have to double check tomorrow.

 

Side note.  One of the major holdups on beta15 has been a bug we've been chasing down relating to btrfs.  The bug was particularly nasty because if your system lost power while an open file (vdisk in our case) existed on the device formatted with btrfs, upon reboot, the array would hang at mount.  You could SSH in and see the command stuck, but could do nothing to kill it.  Tonight that bug has been averted by upgrading to the recently released Linux 4.0 kernel.  The 4.0 kernel is, in Torvalds words: 

 

Linux 4.0 was a pretty small release both in linux-next and in final

size, although obviously "small" is all relative. It's still over 10k

non-merge commits. But we've definitely had bigger releases (and

judging by linux-next v4.1 is going to be one of the bigger ones).

 

Which is all good. It definitely matches the "v4.0 is supposed to be a

_stable_ release", and very much not about new experimental features

etc.

(from https://lkml.org/lkml/2015/4/12/178)

 

Anyhow, just thought I'd chime in with the good news since I just ran the first test on this bug with the new build and it seems okay so far.  Keep fingers crossed!

Link to comment

 

My bad.  So I guess it affects Ivy as well then?  I thought we tested an Ivy system here and didn't see the issue, but will have to double check tomorrow.

 

My system is an Ivy Bridge i3-3220T and it has not properly throttled since I upgraded to 14b from 5.0.6 about a month ago. It definitely affects Ivy Brudge as well.

Link to comment

My system is an Ivy Bridge i3-3220T and it has not properly throttled since I upgraded to 14b from 5.0.6 about a month ago. It definitely affects Ivy Brudge as well.

 

I have Xeon E3-1220 v2 in my system and I can also confirm the Ivy is affected.

 

Tested this on an Ivy Bridge system (non-Xeon) that we have here and scaling works fine.  Not sure what differences exist between an i5 ivy bridge and xeon ivy bridge that would cause the the scaling issue to occur on one and not the other, but that is the case.

Link to comment

My system is an Ivy Bridge i3-3220T and it has not properly throttled since I upgraded to 14b from 5.0.6 about a month ago. It definitely affects Ivy Brudge as well.

 

I have Xeon E3-1220 v2 in my system and I can also confirm the Ivy is affected.

 

Tested this on an Ivy Bridge system (non-Xeon) that we have here and scaling works fine.  Not sure what differences exist between an i5 ivy bridge and xeon ivy bridge that would cause the the scaling issue to occur on one and not the other, but that is the case.

 

perhaps it's in the different technologies used for speed stepping between server and desktop processors:

 

Intel® Demand Based Switching is a power-management technology in which the applied voltage and clock speed of a microprocessor are kept at the minimum necessary levels until more processing power is required. This technology was introduced as Intel SpeedStep® Technology in the server marketplace.
Link to comment

Tested this on an Ivy Bridge system (non-Xeon) that we have here and scaling works fine.  Not sure what differences exist between an i5 ivy bridge and xeon ivy bridge that would cause the the scaling issue to occur on one and not the other, but that is the case.

 

Hi jonp,

Please find a screenshot for Xeon E3 1220 V2

Just to clarify that this "bug" is not making me nervous at all.

By the way noticed it a week ago...

 

Screen_Shot_2015-04-15_at_7_15.55_PM.png.985fec69e329a819de7020600b8ba70f.png

Link to comment

Just an FYI, we're probably HOURS from the next release of beta15 ;-)

 

OK, so throw us a bone here.....

 

What are the changes/updates?

 

Here's the changelog up through 14g (internal release).  The only thing not documented below that I know is happening for 15 is the upgrade to Linux Kernel 4.0.

 

Changes

=======

 

Version 6.0-beta14g (internal release)

-------------------

- docker: update to 1.5.0

- libvirt: update to 1.2.13

- php: update to 5.4.39

- qemu: update to 2.2.1

- slack: updated usb.ids and pci.ids files

- slackware Security Avisory updates: 

  samba: 4.1.17 

  sudo: 1.8.12

- webGui: include dynamix.vm.manager

 

Version 6.0-beta14f (internal release)

-------------------

- btrfs-progs: update to 3.19

- emhttp: more btrfs pool operation refinements

- linux: update to 3.19.3

- linux: set CONFIG_PREEMPT: Preemptible Kernel (to address RCU timeout errors)

- mover: delete partial destination files on error

- unraid: eliminate mdcmd semaphore; rename certain disk status codes

 

Version 6.0-beta14e (internal release)

-------------------

- emhttp: improve btrfs pool device operations

- slack: invoke '/usr/local/sbin/notify init' within rc.local to initialize notification subsystem.

- webGui: dynamix 2015.03.17

- webGui: apcupsd plugin: remove delays per dlandon.

- webGui: add explanatory text if device is unmountable

- webGui: support "file system check" and "btrfs balance" background operation

 

Version 6.0-beta14d (internal release)

-------------------

- Less verbose plugin install progress and messages.

- Eliminate the ”no-check-certificate" option to wget in the plugin download function.  This is because 

  we DO have a valid SSL cert and we DO want it checked when someone downloads unraid.  For 3rd party plugins 

  that bring code from non-trusted source, they have to put that switch on the URL inside the plugin. 

- Fixed that Cache Devices improperly disappeared from Main page for certain cases.

 

Version 6.0-beta14c (internal release)

-------------------

- More device spinning/temperature refinements.

Link to comment

 

Some light reading material for you while you wait:  http://lime-technology.com/wiki/index.php/UnRAID_Manual_6

 

Note:  the above is a work in progress.

Just a thought - I would have put the section on the GUI (Section 8) ahead of the section on Docker (currently section 5).  Not sure why but that seems a more logical order for me.  The rationale is that all users use the GUI, but only some use Dockers.

Link to comment

 

My Ivy Bridge i5-3570S throttles OK.  Turbos OK too.

 

How does one determine if there cpu is throttling correctly? I have no idea what mine should or should not look like?

 

If you don't know how to determine it, then you probably shouldn't care about it ;-).

Link to comment

 

Some light reading material for you while you wait:  http://lime-technology.com/wiki/index.php/UnRAID_Manual_6

 

Note:  the above is a work in progress.

Just a thought - I would have put the section on the GUI (Section 8) ahead of the section on Docker (currently section 5).  Not sure why but that seems a more logical order for me.  The rationale is that all users use the GUI, but only some use Dockers.

 

The entire GUI section is mostly a copy+paste from unRAID 5's wiki, so it's needing a refresh as well and as far as chronological order goes, I agree that when it's refreshed, it should sit higher in the order of things.  To us, there are three layers to unRAID:

 

1)  Software-defined NAS

2)  Application Server

3)  Virtualization Host

 

The way the wiki should be explained is also in that order.  The NAS functionality is pretty much everything EXCEPT for Docker + VMs, so all that documentation belongs ABOVE the documentation for Docker/VMs.  I guess this was a long long long way of writing:  "I agree with you." ;-)

Link to comment

 

My Ivy Bridge i5-3570S throttles OK.  Turbos OK too.

 

How does one determine if there cpu is throttling correctly? I have no idea what mine should or should not look like?

 

I think you look at the main dashboard and in the bottom left corner, check the speed.  Apparently should throttle down, I think it's just an issue for Haswell.

Link to comment

 

 

My Ivy Bridge i5-3570S throttles OK.  Turbos OK too.

 

How does one determine if there cpu is throttling correctly? I have no idea what mine should or should not look like?

 

If you don't know how to determine it, then you probably shouldn't care about it ;-).

 

That's exactly what I have been doing :D every time I saw something about it I just said meh... But after the page long conversation here I got curious...

Link to comment

 

 

My Ivy Bridge i5-3570S throttles OK.  Turbos OK too.

 

How does one determine if there cpu is throttling correctly? I have no idea what mine should or should not look like?

 

If you don't know how to determine it, then you probably shouldn't care about it ;-).

 

That's exactly what I have been doing :D every time I saw something about it I just said meh... But after the page long conversation here I got curious...

 

You will see the speeds drop significantly. Mine go down to about 800 I think, and then all the way up to 3500

Link to comment

We're really excited to get this release out.  Not only do we have a ton of bug fixes and general package updates, the VM manager is going to be a HUGE upgrade in terms of functionality.  A huge thanks to dmacias here for his work on the initial plugin and eschultz for taking it to the next level with dynamic webGui controls for adding multiple devices, support for GUI-driven host-device assignments, CPU pass through / pinning, and much more.  The documentation in the wiki says a lot, but a picture is worth a thousand words:

 

vm-manager-beta15.jpg

 

Eric's work on the presentation side of this is really top-notch.  Using the controls on the left of vDisks, Graphics, Sound, and Network, you can add multiple devices.  Eric also put good logic into form to ensure optimal settings are automatically applied.  Examples: 

  • If you assign an NVIDIA GPU to a Windows VM, the Hyper-V settings are automatically disabled.
  • If you create a non-Windows VM, the Q35 chipset is automatically selected for you.
  • You can have VNC graphics for the primary graphics device, but not secondary devices.
  • When selecting CPUs to pin to, we prevent you from unchecking all the options (must have at least one CPU core assigned).

 

We're doing our best to make this as simple as possible.  In fact, that screenshot is of the "advanced view" display.  The basic view hides a lot of those controls as many are not necessary for the average user to tweak.

Link to comment

I should also mention that another big upgrade from the previous plugin version is the ability to edit VMs using a GUI editor like when creating.  The prior plugin didn't offer a GUI based method to editing VMs.  Instead, you would edit the XML for them directly (which we still support and with a new and improved built-in editor).  This was actually a pretty huge undertaking by Eric and it works very well.

Link to comment

I should also mention that another big upgrade from the previous plugin version is the ability to edit VMs using a GUI editor like when creating.  The prior plugin didn't offer a GUI based method to editing VMs.  Instead, you would edit the XML for them directly (which we still support and with a new and improved built-in editor).  This was actually a pretty huge undertaking by Eric and it works very well.

 

I'm just editing a VM xml.  Think I'll wait.  This looks absolutely awesome!!

Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.