unRAID OS version 6.3 is primarily a bug fix and package "catch up" release.  There is one interesting new feature:

• IGD Device Assignment to VMs
  

 

Note: Please remove any custom "vfs fruit" variables you may have added under Settings/SMB Samba extra configuration section since those are already enabled by default since rc5.

 

Please refer to change log below for list of updated components.

 

unRAID Server OS Change Log
===========================

Version 6.3.0-rc6 2016-12-07
----------------------------

Base distro:

- inotify-tools: version 3.14
- mozilla-firefox: version 50.0.2 (CVE-2016-9078, CVE-2016-9079)
- ntfs-3g: version 2016.2.22
- rsyslog: version 8.23.0

Linux kernel:

- version 4.8.12 (CVE-2016-8650)

Management:

- network: must set forwarding delay before setting stp 'on'; need 1 sec delay between stop/start in restart
- samba: move OS X "vfs_fruit" optimization from 'global' to individual share settings
- shutdown: increase default shutdown delay from 60 to 90 sec; eliminate extraneous console messages
- update: hwdata/{pci.ids,usb.ids,oui.txt,manuf.txt} smartmontools/drivedb.h
- webGui: add Hibernate option to VM context menu

Version 6.3.0-rc5 2016-11-25
----------------------------

Base distro:

- bash: version 4.4.005
- grep: version 2.26 (rev2)
- kernel-firmware: version 20161118git
- libvirt: version 2.4.0
- mozilla-firefox: version 50.0
- ntp: version 4.2.8p9 (CVE-2016-9311, CVE-2016-9310, CVE-2016-7427, CVE-2016-7428, CVE-2016-9312, CVE-2016-7431, CVE-2016-7434, CVE-2016-7429, CVE-2016-7426, CVE-2016-7433)

Linux kernel:

- version 4.8.10 (CVE-2016-7039)

Management:

- add btrfs_check script to facilitate btrfs file system check
- bug fix: firefox correctly connects to emhttp port number
- samba: include OS X vfs_fruit optimizations by default
- shutdown: bug fix: overall shutdown timeout read from wrong cfg file
- shutdown: bug fix: saving shutdown timeout value
- shutdown: add config setting for overall VM shutdown timeout
- update: hwdata/{pci.ids,usb.ids,oui.txt,manuf.txt} smartmontools/drivedb.h
- webGui: add filesystem Check for btrfs volumes in Maintenance mode
- webGui: add vm shutdown time-out option
- webGui: generate VM xml with cpu topology as single-threaded when host cpu is AMD
- webGui: special text aside new disks to clear
- webGui: split out Europe/Istanbul as its own timezone

Other:

- make_bootable_linux: use readlink instead of realpath since it is more compatible

Version 6.3.0-rc4 2016-11-11
----------------------------

Base distro:

- curl: version 7.51.0 (CVE-2016-8615, CVE-2016-8616, CVE-2016-8617, CVE-2016-8618, CVE-2016-8619, CVE-2016-8620, CVE-2016-8621, CVE-2016-8622, CVE-2016-8623, CVE-2016-8624, CVE-2016-86$
- docker: version 1.12.3 (CVE-2016-8867)
- glibc-zoneinfo: version 2016i
- grep: version 2.26
- kernel-firmware: version 20161001git
- libusb: version 1.0.21
- libX11: version 1.6.4 (CVE-2016-7942, CVE-2016-7943)
- libXfixes: version 5.0.3 (CVE-2016-7944)
- libXi: version 1.7.8 (CVE-2016-7945, CVE-2016-7946)
- libXrandr: version 1.5.1 (CVE-2016-7947, CVE-2016-7948)
- libXrender: version 0.9.10 (CVE-2016-7949, CVE-2016-7950)
- libXtst: version 1.2.3 (CVE-2016-7951, CVE-2016-7952)
- libXv: version 1.0.11 (CVE-2016-5407)
- lvm2: version 2.02.166
- mozilla-firefox: version 49.0.2
- mpfr: version 3.1.5
- nano: version 2.7.1
- php: version 7.0.13
- samba: version 4.5.1

Linux kernel:

- version 4.8.7 (CVE-2016-7042, CVE-2016-8630, CVE-2016-8633)
- md/unraid version: 2.7.1
  - make the 'check' command "correct"/"nocorrect" argument case insensitive
  - mark superblock 'clean' upon initialization

Management:

- emhttp: add API command cmdStopMover and API status variable shareMoverActive (yes/no)
- emhttp: ensure disk shares have proper permissions set even if not being exported
- emhttp: fix detecton of unclean shutdown to trigger automatic parity check upon Start if necessary
- emhttp: fix mac address not set in avahi afp share advertisement (regression introduced in -rc2)
- emhttp: unmount docker/libvirt loopback if docker/libvirt fail to start properly
- mover: add start/stop/status subcommands (experimental)
- mover: dereference files on user shares to get st_dev/st_ino instead of from shfs stat structure
- shfs: eliminate 'use_ino' option dependency (fixes NFS issue on buggy clients)
- update: hwdata/{pci.ids,usb.ids,oui.txt,manuf.txt} smartmontools/drivedb.h
- webGui: combine PCI Devices and IOMMU Groups on Sysdev page; Sort IOMMU and USB Devices properly
- webGui: correct button positions on confirmation dialogs
- webGui: correct handling of unclean shutdown detection
- webGui: fixed device passthrough for Docker
- webGui: fixed Docker/VMs tab stay visible when array is stopped
- webGui: fixed some help text typos
- webGui: update to font-awesome 4.7.0

Version 6.3.0-rc3 2016-10-22
----------------------------

Linux kernel:

- version 4.8.4 (CVE-2016-5195)

Management:

- shutdown: additional logging in event of cmdStop timeout

Version 6.3.0-rc2 2016-10-19
----------------------------

Base distro:

- docker: version 1.12.2
- containerd: version 0.2.4
- runc: version 1.0.0_rc2
- libseccomp: version 2.3.1
- libvirt-php: 0.5.2 (+ latest from git)
- php: version 7.0.12
- usbredir: version 0.7.1
- spice: version 0.12.8
- spice-protocol: version 0.12.12
- imlib2: version 1.4.9
- gd: version 2.2.3
- rsyslog: version 8.22.0
- libfastjson: version 0.99.4
- reverted OVMF firmware back to the version in 6.2 due to user reported issues

Linux kernel:

- version 4.8.2
- r750: version 1.2.10-16_10_13

Management:

- bug fix: Samba listening on the wrong interface in a VLAN-enable config
- emhttp: added starting/stopping docker/libvirt events
- emhttp: removed 'reboot', 'shutdown', 'changeNetwork' API calls; eliminated emhttp-managed network vars
- md/unraid: kernel 4.8 API changes; support FUA and SYNC write operation flags
- shutdown: save diagnostics in event of cmdStop timeout
- vm manager: add 'virtio-win-0.1.126-2' and 'virtio-win-0.1.126-1' to VirtIO-ISOs list
- webGui: Added "Join" notification agent
- webGui: Add support for "Notify My Android" notifications
- webGui: Fixed: Windows unable to extract diagnostics zip file
- webGui: Increased password length to 128 characters
- webGui: Interactive Reboot and Shutdown directly invokes '/sbin/reboot' and '/sbin/poweroff'

Version 6.3.0-rc1 2016-10-05
----------------------------

Base distro (primary packages):

- btrfs-progs: version v4.7.2
- netatalk: version 3.1.10
- openssl: version 1.0.2j
- php: version 5.6.26
- qemu: version 2.7.0
- samba: version 4.5.0

Base distro (secondary packages):

- apcupsd: version 3.14.14
- bash: version 4.4.0
- curl: version 7.50.3
- diffutils: version 3.5
- e2fsprogs: version 1.43.3
- etc: version 14.2
- file: version 5.28
- fuse: version 2.9.7
- gawk: version 4.1.4
- gdbm: version 1.12
- glib2: version 2.46.2
- glibc: version 2.24
- glibc: version zoneinfo
- glibc-solibs: version 2.24
- gnutls: version 3.4.15
- jemalloc: version 4.2.1
- kmod: version 23
- libtasn1: version 4.9
- libxml2: version 2.9.4
- lvm2: version 2.02.161
- nano: version 2.7.0
- network-scripts: version 14.2
- nfs-utils: version 1.3.4
- openssl-solibs version: 1.0.2j
- pkgtools-14.2-noarch: version 13
- reiserfsprogs: version 3.6.25
- util-linux: version 2.28.2

Base distro (desktop GUI packages):

- freetype: version 2.6.5
- harfbuzz: version 1.3.0
- libXfont: version 1.5.2
- mesa: version 12.0.2
- mozilla-firefox: version 49.0.1
- xkeyboard-config: version 2.18
- xorg-server: version 1.18.4

Linux kernel:

- version 4.7.6
- aacraid: version 1.2.1-52011
- intel 10Gbit ethernet: revert to in-tree drivers
- r750: version 1.2.9-16_09_27

Management:

- bug fix: For file system type "auto", only attempt btrfs,xfs,reiserfs mounts.
- bug fix: For docker.img and libvirt.img, if path on /mnt/ check for mountpoint on any subdir component
- bug fix: During shutdown force continue if array stop taking too long.
- bug fix: Handle case in 'mover' where rsync may move a file but also return error status.
- md/unraid: Fix bug where case of no data disks improperly detected.
- webGui: Add "Shutdown time-out" control on Disk Settings page.
- webGui: bug fix: Cannot manually specify virtIO iso in VM settings
- webGui: other changes: refer to github commit log

Upgraded no problem. nvme drive temps are still not showing though 

Upgrade went Smooth, I received the following error from Fix common problems after reboot:

 

Share cache is identically named to a disk share

UPDATE: I brought my VM back online after I disabled my USB 3.1 + Ethernet card that was stubbed through.

 

ID: VIA Technologies VL805 USB 3.0 Host Controller | USB controller (07:00.0) ID = 1106:3483

 

This stinks because this is my card for the HTC Vive I JUST got working this past weekend.

 

Excerpt from my syslinux.cfg for my flash.

label unRAID OS GUI Mode
  menu default
  kernel /bzimage
  append pci-stub.ids=1106:3483,8086:8d26,8086:8d2d,1b21:1242 initrd=/bzroot,/bzroot-gui

 

 

Original post:

This killed my Windows 10 VM running Q35-2.4 with a custom VirtIO raw image on my NVMe.

 

They now hang at the OS booting screen.

 

 

 

This is the log from the VM, with the error on the last line:

 

2016-12-07 18:28:59.243+0000: starting up libvirt version: 2.4.0, qemu version: 2.7.0, hostname: Tower
LC_ALL=C PATH=/bin:/sbin:/usr/bin:/usr/sbin HOME=/ QEMU_AUDIO_DRV=none /usr/local/sbin/qemu -name guest=STEAM-WIN10,debug-threads=on -S -object secret,id=masterKey0,format=raw,file=/var/lib/libvirt/qemu/domain-5-STEAM-WIN10/master-key.aes -machine pc-q35-2.4,accel=kvm,usb=off,dump-guest-core=off,mem-merge=off -cpu host -drive file=/usr/share/qemu/ovmf-x64/OVMF_CODE-pure-efi.fd,if=pflash,format=raw,unit=0,readonly=on -drive file=/etc/libvirt/qemu/nvram/598f30c6-3d2b-1f58-f6ca-695c0e3ca193_VARS-pure-efi.fd,if=pflash,format=raw,unit=1 -m 16384 -realtime mlock=off -smp 8,sockets=1,cores=4,threads=2 -uuid 598f30c6-3d2b-1f58-f6ca-695c0e3ca193 -display none -no-user-config -nodefaults -chardev socket,id=charmonitor,path=/var/lib/libvirt/qemu/domain-5-STEAM-WIN10/monitor.sock,server,nowait -mon chardev=charmonitor,id=monitor,mode=control -rtc base=localtime,driftfix=slew -global kvm-pit.lost_tick_policy=discard -no-hpet -no-shutdown -boot strict=on -device i82801b11-bridge,id=pci.1,annel0,path=/var/lib/libvirt/qemu/channel/target/domain-5-STEAM-WIN10/org.qemu.guest_agent.0,server,nowait -device virtserialport,bus=virtio-serial0.0,nr=1,chardev=charchannel0,id=channel0,name=org.qemu.guest_agent.0 -device vfio-pci,host=03:00.0,id=hostdev0,bus=pci.2,addr=0x4 -device vfio-pci,host=03:00.1,id=hostdev1,bus=pci.2,addr=0x5 -device vfio-pci,host=07:00.0,id=hostdev2,bus=pci.2,addr=0x6 -device usb-host,hostbus=1,hostaddr=5,id=hostdev3,bus=usb.0,port=1 -device virtio-balloon-pci,id=balloon0,bus=pci.2,addr=0x7 -msg timestamp=on
Domain id=5 is tainted: high-privileges
Domain id=5 is tainted: host-cpu
char device redirected to /dev/pts/0 (label charserial0)
2016-12-07T18:29:01.371672Z qemu-system-x86_64: warning: Unknown firmware file in legacy mode: etc/msr_feature_control

Upgrade went Smooth, I received the following error from Fix common problems after reboot:

 

Share cache is identically named to a disk share

That is telling you that you have a user share named "cache". Any top level folder on cache or any array drive is automatically a user share. You probably accidentally created a folder named "cache" by misconfiguring some application. Figure out what is in it, get rid of it, and fix whatever caused it.

Upgrade went Smooth, I received the following error from Fix common problems after reboot:

 

Share cache is identically named to a disk share

Are you sure you don't have a share named "cache"?  This will tell the whole story:

 

ls /mnt/user

 

Admitedly, I haven't upgraded yet as I'm at work

There was one there, looks like in my install of teamspeak I didnt set the location properly!!

After updating to rc5 I noticed a problem with my Ubuntu server VM (16.04 LTS.) A share I have on my cache drive (I have a 2 disk cache pool Raid 1)which is mounted in the VM  is now not writable which I discovered when my SABnzbd running within the VM started throwing off error messages when trying to create directories within the share.  I checked the permissions, they are ok.  I tried creating another share on the cache drive, same result.  I scrubbed the cache pool and no errors.  All other shares which are on the array (not on the cache drive) are behaving normally.  Any suggestions?

 

Updated to rc-6, same problem as previously reported after updating to rc-5.

Check the mount point for the share inside the VM, and which user mounted it? I've found enough times that only the user that mounted it can write to it.

Check the mount point for the share inside the VM, and which user mounted it? I've found enough times that only the user that mounted it can write to it.

 

I mount the share through fstab using 9p.  The permissions of fstab are 644 i tried changing permissions to 777 (which is probably not a good idea), same result.

- samba: move OS X "vfs_fruit" optimization from 'global' to individual share settings

 

Sorry, where is this setting?  I don't see it under "shares" anywhere.  Or do I need to enable something globally first?

 

Also, the man page for vfs_fruit says:

Be careful when mixing shares with and without vfs_fruit. OS X clients negotiate SMB2 AAPL protocol extensions on the first tcon, so mixing shares with and without fruit will globally disable AAPL if the first tcon is without fruit.

 

Maybe this should be mentioned in the release notes or the help?

- samba: move OS X "vfs_fruit" optimization from 'global' to individual share settings

 

Sorry, where is this setting?  I don't see it under "shares" anywhere.  Or do I need to enable something globally first?

It's not configurable.  Those lines are included in every share section except the flash share - that was the fix.

 

Also, the man page for vfs_fruit says:

Be careful when mixing shares with and without vfs_fruit. OS X clients negotiate SMB2 AAPL protocol extensions on the first tcon, so mixing shares with and without fruit will globally disable AAPL if the first tcon is without fruit.

 

Maybe this should be mentioned in the release notes or the help?

That was the reason for adding those lines for every share, but sure, if someone accesses 'flash' share first maybe above will trigger.  Not sure what would make it "un-trigger".

Ah, thanks for that. Sounds good!

 

Sent from my ONEPLUS A3000 using Tapatalk

 

 

Have been having some severe problems with moving files from my Windows 10 PC to my unraid Array or Cache drive(Not related to outside post). They would begin to transfer and then pause for minutes at a time. I have encountered this with significantly less severity since before RC6(30 second pauses randomly on several gig moves, maybe increasing transfer time by 1.5x-2x), but with RC6 it was more like 100x, at least when it was many smaller files. Strangely last night, a several gig movie managed to copy in only like 5x the usual time.

 

Anyway, I just reverted to RC5 and was able to complete a transfer that I had been fighting all morning without any interruption.

Next time you prepare such activity, open an SSH terminal to your unRAID machine, and install atop and run it.

 

atop has been added to NerdPack. Use that to install it, and run it from ssh to look for bottlenecks.

Next time you prepare such activity, open an SSH terminal to your unRAID machine, and install atop and run it.

 

1) Download atop-2.2-x86_64-1_SBo.tgz somewhere in your server.

2) Run installpkg on it.

3) Run atop.

4) Watch for blinking red status lines.

 

Hmm, maybe this package could be added to the NerdPack. I'll propose the author to add this tool to his build list.

Done

Next time you prepare such activity, open an SSH terminal to your unRAID machine, and install atop and run it.

 

1) Download atop-2.2-x86_64-1_SBo.tgz somewhere in your server.

2) Run installpkg on it.

3) Run atop.

4) Watch for blinking red status lines.

 

Hmm, maybe this package could be added to the NerdPack. I'll propose the author to add this tool to his build list.

Done

 

"Done" as in that package added to nerd pack?  Bear in mind that unless you compile the code yourself and verify the source is unaltered, or download from a known trusted source, you really don't know what is being installed.  I'm not making any judgement on that particular package and I have no reason to suspect it's anything other than it purports to be... just saying.  In general this is the major problem with plugins: they run as 'root' and it would be ridiculously easy for malware to be hiding in there.

 

With all the attention we've been placing on security, install such packages at your own risk.

Next time you prepare such activity, open an SSH terminal to your unRAID machine, and install atop and run it.

 

1) Download atop-2.2-x86_64-1_SBo.tgz somewhere in your server.

2) Run installpkg on it.

3) Run atop.

4) Watch for blinking red status lines.

 

Hmm, maybe this package could be added to the NerdPack. I'll propose the author to add this tool to his build list.

Done

 

"Done" as in that package added to nerd pack?  Bear in mind that unless you compile the code yourself and verify the source is unaltered, or download from a known trusted source, you really don't know what is being installed.  I'm not making any judgement on that particular package and I have no reason to suspect it's anything other than it purports to be... just saying.  In general this is the major problem with plugins: they run as 'root' and it would be ridiculously easy for malware to be hiding in there.

 

With all the attention we've been placing on security, install such packages at your own risk.

Sorry, I didn't mean to reply to this thread. And I didn't add that particular package linked above. If the package isn't available on a trusted repo, I usually compile it myself from the original source.

I only supplied it to get the user here going quickly. Naturally, it's not a trusted repository of packages. I assumed that dmacias would build it himself.

 

It has no dependencies, and building it merely requires basic build tools installation on Slackware 14.1, and the atop package from Slackbuilds, if you trust them.

 

I intend to contribute to a "trusted" repository for unRAID-ZFS, at least for alternative branches. My builds are intended to be tested and possibly trusted on SSD block devices, as they are based on spl:master and zfs:ntrim (from dweeezil repository). I have packages for my personal use built for 6.2.4 and 6.3.0-rc6, using the Slackbuilds scripts for spl-solaris and zfs-on-linux, with both modified to include a different version name, and the latter modified to pass --with-spl=/tmp/SBo/spl-<version>.

 

I'm not sure what to do to earn trust for packages, though.

 

E: Removed binary package from my bucket, as NerdPack makes it redundant.

So since going up to rc6 i'm having a weird issue with my EMBY docker filling up with multiple gig log files

 

EMBY was updated around the same time but as i am currently cleaning the logs twice a day i wanted to rule out it being a weird docker issue. I have posted a chunk of the error on the support thread for the docker with no luck, worth a try

 

Attached is a copy of some of the log after i shut down the docker as well as my server diagnostic. This log files can get to around 4GB multiple times a day so its a little frustrating, any help finding out where the issue lies (i suspect in the docker itself) would be great

 

Jamie

log.txt

archangel-diagnostics-20161213-1905.zip

I've dumped the ZFS stuff, probably too crazy to be using that bleeding edge code anyway. And now I'm down one SSD, it may have failed, or the cable or port may have failed. Now I've got one cache drive, the 256GB, formatted XFS.

 

Seems to be in working order now, though.

Anyone come across this error in the log before? nvme passthrough worked fine in 6.2.x.

 

I am passing through my samsung nvme 950 pro drive to a windows 10 VM. It starts up fine on the first boot.

However, when I restart windows or power off/on the VM i get the following error in UNRAID and unable to boot from the nvme drive again until i restart the unraid host. I tried a different slot for the nvme drive and also the vfio_iommu_type1.allow_unsafe_interrupts=1 in sysinux but it doesnt make a difference. The nvme drive is in its own IOMMU group.

 

Dec 17 10:33:55 MOUNRAID01 kernel: vfio-pci 0000:4f:00.0: Failed to return from FLR

Dec 17 10:33:59 MOUNRAID01 kernel: vfio_bar_restore: 0000:4f:00.0 reset recovery - restoring bars

Dec 17 10:33:59 MOUNRAID01 kernel: vfio_bar_restore: 0000:4f:00.0 reset recovery - restoring bars

...

 

EDIT:Could be the same issue as experienced in this thread.. https://lime-technology.com/forum/index.php?topic=36768.msg516497#msg516497

 

diagnostics attached.

 

I also ran lspci -vv and the nvme device shows as being FLR capable.

 

DevCap: MaxPayload 128 bytes, PhantFunc 0, Latency L0s unlimited, L1 unlimited

                        ExtTag- AttnBtn- AttnInd- PwrInd- RBE+ FLReset+ SlotPowerLimit 25.000W

mounraid01-diagnostics-20161217-1326.zip

Anyone come across this error in the log before? nvme passthrough worked fine in 6.2.x.

 

 

I am using 6.3RC6.

 

You should attach your entire diagnostics file if  you are expecting someone to be able to figure out what is going on...

Anyone come across this error in the log before? nvme passthrough worked fine in 6.2.x.

 

 

I am using 6.3RC6.

 

You should attach your entire diagnostics file if  you are expecting someone to be able to figure out what is going on...

 

OK, I have done this. Would appreciate it if someone could take a look.

Another issue that I have seen in all 6.3.RC releases is that

<alias name='virtio-disk2'/> does not stick after updating the VM XML.

Because of this I cannot use qemu args that use the alias name without an error being thrown up.