unRAID Server Release 6.0-beta5a-x86_64 Available


Recommended Posts

  • Replies 148
  • Created
  • Last Reply

Top Posters In This Topic

Have an weird one here.  ASRock MB (890FX Deluxe 4) that has IOMMU and the AMD 1090T processor that I previously ran ESXi on.  I am making the switch to unRaid w/ native Xen.

 

If I boot unRaid, everything works fine. (Syslog attached "unRaid-Boot.txt).  If I boot Xen, I get a whole bunch of failures related to the hard drives and when it finally makes it to the unraid webgui most of my drives are missing. (Syslog attached "Xen-Boot.txt).

 

Any ideas?  Thanks in advanced!

 

Searching some more, I think Xen is maybe reusing IRQs?  This thread (without resolution) on an old XEN mailing list is one of the only references I can find.

 

http://old-list-archives.xenproject.org/archives/html/xen-users/2006-10/msg00289.html

Link to comment

Just successfully finished moving my main server over from 5.0.5 to 6b5a. Also cleaned a lot of files in the process. I'm wondering where midnight commander went in unraid v6. Can anyone please give me directions on how in re-install it? Thanks!

Link to comment

Can anyone please give me directions on how in re-install it? Thanks!

 

Midnight Commander is included by default, both on unraid 5 and 6, just type mc to start it

 

You are absolutely correct! I must be very tired. I remoted in to my esxi server instead of unrad. lol! Thanks!

Link to comment

Have an weird one here.  ASRock MB (890FX Deluxe 4) that has IOMMU and the AMD 1090T processor that I previously ran ESXi on.  I am making the switch to unRaid w/ native Xen.

 

If I boot unRaid, everything works fine. (Syslog attached "unRaid-Boot.txt).  If I boot Xen, I get a whole bunch of failures related to the hard drives and when it finally makes it to the unraid webgui most of my drives are missing. (Syslog attached "Xen-Boot.txt).

 

Any ideas?  Thanks in advanced!

 

Figured it out!  So two of the drives were set to IDE for some reason.  Once I changed them back to ACHI it booted fine.

Link to comment

[sOLVED - run modprobe sg]

 

Tried 6.0-beta5a today, upgrading from 5.0.4, and found SCSI Generic support was lost.  This is necessary for using Smartctl with Areca controllers, since you need the generic SCSI device (/dev/sgX) and not the block device.

 

Run lsscsi -g to get the generic SCSI device IDs  or look in /proc/scsi/sg/device_hdr and /proc/scsi/sg/devices.

 

 

Link to comment

OK So I've sent this directly to Tom as well but i wanted to get the masses involved too.

 

I used to be running the system shown below in my signature with no troubles at all.  then I tried upgrading my hardware to Asus M5A97 R2.0 / FX-8320 / Kingston 8GB ECC and all hell broke loose.  In order:

 

- First thing I noticed was that bzroot was loading VERY slowly.  I played around with the BIOS looking for any thing that might be driving it to USB1.1 but nothing helped.  I did let it boot though and there were still problems which I will talk about below ...

 

- So I copied everything to a new USB, and boom bzroot loads up just as fast as always, so maybe it was one of those crappy flash drives you mentioned in a thread BUT ...

 

- The system is not right regardless.  There are a lot of things all wrong (no ip address, /boot is mostly empty) which culminate in me figuring out that /UNRAID is not mounting.  I'm seeing 30 seconds of "waiting for /dev/disk/by-label/UNRAID (will check for 30 sec)..." only to have it never find UNRAID leaving me with no webgui, and as best i can tell no way to copy a system log. I copy it to /boot and it is gone when i move the USB to my PC, I even copied it to /boot/config (the only directory even showing in the console) and still syslog is not there.

 

- next I load up a 100% fresh clean copy of unraid 6.0b5a onto two different thumbdrives (mushkin Ventura Pro 32gb usb 3 and DanElec 4gb) and I get the exact same problems

 

- Some more viewing of the boot process yields some other errors like, "device not accepting address 2 error -32" as well as "device descriptor read/64, error -32"

 

- So for grins I load up a fresh copy of 5.0.5 and guess what?  It booted perfectly with not even a hiccup.

 

- Just to make sure I didn't somehow change anything possible at all from that last attempt I immediately put 6.0b5a back onto that same usb, with no bios changes, and all my problems come back.

 

Sooo, first, do you know how I can get a log out of this thing when it doesn't actually seem to be saving to the flash when i copy to /boot or /boot/config (remember /boot is otherwise empty at this point too unlike when i mount the flash on my PC)  I also can’t seem to telent in because I’m not getting an IP address per ifconfig.

 

I'm not sure what good it would do since so much flies by so fast, but i made a video of the boot process and I could post it to my gdrive if anyone thought it would help.

 

EDIT1: I also tried booting with and without iommu enabled

 

EDIT2: upgraded bios from 2201 to 2301, not change except now i'm seeing -110 errors vice -32  Can't say that is progress :(

 

EDIT3: [sigh] ok I KNOW I tried this variation on the theme, but I'm willing to accept that maybe I didn't.  Or it was the upgraded BIOS that finally did it BUT I found the problem.  In the IOMMU section of the bios there is another option beyond just enable/disable.  There is also an option for "IOMMU Mode" with the options for disable/64mb (see attached).  Well apparently it must be in disabled mode. 

 

So once I validated that option was THE thing making the difference (several back and forth tests) I then booted into the XEN option to see what it had to say for itself.  A list of PCI Devices shows in attached 2.  I'm not sure if that means I can start passing through devices, but I'm just adding it for completeness [i think hahaha]

iommu.jpg.4c99ddf10683a246c1f4fb6a3eae0315.jpg

pci_devices.png.bcc185626df9baf2424e36f4386902c8.png

Link to comment

.  A list of PCI Devices shows in attached 2.  I'm not sure if that means I can start passing through devices, but I'm just adding it for completeness [i think hahaha]

 

Yup, you are ready except it looks like you only have a nvidia GPU.  No graphics pass through has been working for us with nvidia cards just yet.  Hope you have an AMD GPU around ;-)

 

Sent from my Nexus 5 using Tapatalk

 

 

Link to comment

Well I'm not really looking to pass anything thru at this point in time, but i wanted the option in the future.  The nvidia is just a place holder (pulled from my pc) because this MB didn't have an onboard gpu like my old MB did.  That said, I have something likely just as useless on its way (or not?) an  ATI Rage XL 8MB PCI Video Card.  I got it because it is PCI to save me a PCIe slot.  It is also low power, but that only came into play when deciding between that and another ATI blah blah blah 128 something or other. 

 

In any case it looks like my JetFlash will need to be replaced because it still boots very slow on this MB.  Very strange.  I'm also worried that the "IOMMU Mode" might still be needed for actual passthru (as in set to 64mb) and I wonder if it isn't still a problem / missing [something] in 6.0 because I do know 100% that when it was set to 64mb 5.05 booted perfectly.  But given the description about how it relates to LINUX and converting 32bit i/o to 64 bit I am probably not surprised that 5.0.5 didn't experience any problems [shrug]

Link to comment

THIS http://ubuntuforums.org/showthread.php?t=2114055 described my problem as well.  Differen't MB and same chipset.  And this guy went even further in testing other OS's than my simple 32bit 64bit unraid test.

 

[sheesh] I take that back.  He seems to be fixing his problems by enabling IOMMU.  For me the first setting (enabled|disabled) has no effect.  It is the second setting (disabled|64mb) that works|doesn't work :(  Wonder if it is a reversed setting error in the bios?!?

Link to comment

THIS http://ubuntuforums.org/showthread.php?t=2114055 described my problem as well.  Differen't MB and same chipset.  And this guy went even further in testing other OS's than my simple 32bit 64bit unraid test.

 

[sheesh] I take that back.  He seems to be fixing his problems by enabling IOMMU.  For me the first setting (enabled|disabled) has no effect.  It is the second setting (disabled|64mb) that works|doesn't work :(  Wonder if it is a reversed setting error in the bios?!?

 

Is there an update available for your BIOS?

 

Sent from my Nexus 5 using Tapatalk

 

 

Link to comment

Anybody have any idea why my shares wouldn't show up when upgrading to version 5 to this version?  The array is started and the disk information looks correct.  There .cfg files are there from my previous install, but no shares. If I revert back to v5 everything is fine.

 

EDIT: Nevermind. The array was started in maintenance mode.  I'm and idiot.  Carry on....

Link to comment

woot! Fingers Crossed.  ummm errr ... ETA?  Days? Weeks? Months

 

Now if I could just figure out why this is happening:

 

Production Jetflash 2GB (reg'ed with Limetech): Loads bzroot Slow

Cheap P.O.S. DanElec 4GB: Loads bzroot Fast

Mushkin Ventura Pro 32GB USB3: Loads bzroot Fast 

Brand new SanDisk Cruzer Fit #1: Loads bzroot Slow

Brand new SanDisk Cruzer Fit #2: Loads bzroot Slow

 

I know I could just use the two "fast" ones but the Mushkin isn't cheap and the DanElec doesn't inspire confidence.  That's why i went with the Limetech recommended Sandisks.

 

I know it is not an unraid problem, I'm just venting. 

Link to comment

 

woot! Fingers Crossed.  ummm errr ... ETA?  Days? Weeks? Months

 

Now if I could just figure out why this is happening:

 

Production Jetflash 2GB (reg'ed with Limetech): Loads bzroot Slow

Cheap P.O.S. DanElec 4GB: Loads bzroot Fast

Mushkin Ventura Pro 32GB USB3: Loads bzroot Fast 

Brand new SanDisk Cruzer Fit #1: Loads bzroot Slow

Brand new SanDisk Cruzer Fit #2: Loads bzroot Slow

 

I know I could just use the two "fast" ones but the Mushkin isn't cheap and the DanElec doesn't inspire confidence.  That's why i went with the Limetech recommended Sandisks.

 

I know it is not an unraid problem, I'm just venting.

Slow but reliable and ultra small making braking it off by mistake not possible.

Usb 3 is much faster and Sandisk should be coming out with a usb 3 version soon. But that I believe is why. Remember at the end of the day it is your choice what you get.

 

 

 

Thornwood

Link to comment

OK So I've sent this directly to Tom as well but i wanted to get the masses involved too.

 

...

 

 

I encountered this same issue when working on a new server on the weekend trying to get 6.0-beta5a up and running. 

 

I believe what was happening in my case was that the BIOS would detect the USB key and begin booting unraid, but after the unraid kernel takes over, it was unable to find and mount the USB drive.  This would only occur when I connected the USB drive to USB ports connected to one particular USB controller on my motherboard (GA-Z77X-UD5H), and it would boot fine from any port on the other two controllers.  Mouse and keyboard connected to the rogue USB controller also did not work in unraid. 

 

The USB controller that was causing the problem actually has VIA USB hubs connected to them (on the motherboard) as port expanders - perhaps for some reason unraid was not able to deal with these hubs. 

 

I can replicate this boot failure by hiding a USB controller from unraid using pciback in the sysconfig file, then connecting the boot drive to the hidden controller.

 

When I get some time I'll check the BIOS to see if there are any settings I missed related to IOMMU...  And also cross my fingers for beta6.

Link to comment

Slow but reliable and ultra small making braking it off by mistake not possible.

Usb 3 is much faster and Sandisk should be coming out with a usb 3 version soon. But that I believe is why. Remember at the end of the day it is your choice what you get.

 

Thornwood

 

Slow in this case is not reliable because it speaks to an underlying incompatibility given that other flashdrives are working correctly.  It also causes frustration which is never safe for computer components physical well being ;)  Also since my USB is mounted internal to the case the chances of brake-off are slim to none.  Though I agree in principle and that is why I picked the Fit out of the three Tom recommends.  Well that and i boycott all Sony products reflexively.

 

USB3 is not supported in unraid nor in bios boot so it won't make things any faster.  I know because my USB3 flash boots no faster than the other, USB2.0, flash that also boots quickly.

 

My problem is I'm getting USB1.1 speeds with my OG flashdrive which previously saw normal USB2.0 speeds with my old motherboard.  My two other flashdrives DO get normal speeds with the new motherboard.

 

As luck would have it I picked up a PNY Classic Attache 8GB and it loads lickidy split! :)  So I've emailed Tom for a license transfer.

Link to comment

Well crap Asus JUST uploaded a new bios YESTERDAY for my motherboard AND it appears they have removed the 2301 I upgraded to.  2201, the one I bought the board with, is still showing [sigh]  2301 didn't give me any new problems that weren't there with 2201, but if 2501 fixes that slow flashdrive problem (even though my new PNY works just fine) this will be me http://www.millsworks.net/blog/wp-content/uploads/2009/04/writing_process.gif

 

Of course my real hope is that it makes my IOMMU behavior more logical.  I really should test to see if I can pass something through before and after.

 

Yes, headbanging will commence ... I didn't even read the description because usually all it says is "improve system stability" but nope, this time it says, "Enhance compatibility with some USB device"  :o

 

UPDATE: Nope, no difference.  same flashes are fast and same are slow.  Didn't muck with iommu.

Link to comment

Hi All,

 

Was testing 6.0-beta5a out this week and stumbled upon an issue which I think is not right. Herewith the scenario:

 

- Used the free version for tests. Disk1 and Disk2 shares was enabled and set at the default "High-Water". Disk1 became full with 3G left. Disk2 had 45G left since I deleted some stuff to make space available. I then copied a 5G file to the Data share and it always fail with "(28) No space left on device" from the log. Share is setup to include all disk and exclude none, then.

- I changed the setting to "Most-Free" for the Data share applied it and when copying the 5G file over it failed as above again, then.

- I rebooted the server to make sure the setting is applied, but it failed again. It always start at disk1 to be filled and never copies to disk2 which has the "most-free" space 45G.

 

Is this a bug or setting mismatch?

 

Duppie

Link to comment

Do you have any exclude / include settings applied to your shares?

 

What is your split level set at and what does your folder structure look like?  If you don't have that set correctly then unraid will honor that first and not move on to another drive regardless of your allocation method.

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.