• Unraid OS version 6.8.0-rc9 available


    limetech

    A fairly large kernel update was published yesterday as was a new WireGuard release and I thought it important to update and sanity-check.  We'll let this bake over the weekend and if no new big issue shows up we'll publish to stable. Then we can move on to the 5.4 kernel in Unraid 6.9.

     

    -rc9 summary:

    • Update kernel to 4.19.88.
    • Update to latest WireGuard release
    • Other package updates.

     

    Specific changes in [-rcN] are indicated in bold below.

     

    New in Unraid OS 6.8 release:

     

    The unRAIDServer.plg file (update OS) still downloads the new release zip file to RAM but then extracts directly to USB flash boot device.  You will probably notice a slight difference in speed of extract messages. [-rc2] The 'sync' command at the end has been replaced with 'sync -f /boot'.

     

    Forms based authentication
    If you have set a root password for your server, when accessing webGUI you'll now see a nice login form.  There still is only one user for Unraid so for username enter root.  This form should be compatible with all major password managers out there.  We always recommend using a strong password.  [-rc2] There is no auto-logout implemented yet, please click Logout on menu bar or completely close your browser to logout.

     

    Linux kernel

    • [-rc8] Remains on 4.19
    • [-rc6/-rc7] include latest Intel microcode for yet another hardware vulnerability mitigation.
    • default scheduler now 'mq-deadline' [-rc2] but this can be changed via Settings/Disk Settings/Scheduler setting.
    • enabled Huge Page support, though no UI control yet
    • binfmt_misc support
    • added "Vega 10 Reset bug" [-rc2] and 'navi-reset' patches removed [-rc5]
    • [-rc2] added oot: Realtek r8125: version 9.002.02
    • [-rc3] additional md/unraid changes and instrumentation
    • [-rc6] fix chelsio missing firmware
    • [-rc8] removed Highpoint r750 driver [does not work]

     

    md/unraid driver

     

    Introduced "multi-stream" support:

     

    • Reads on devices which are not being written should run at full speed.  In addition, if you have set the md_write_method tunable to "reconstruct write", then while writing, if any read streams are detected, the write method is switched to "read/modifywrite".
    • Parity sync/check should run at full speed by default.
    • Parity sync/check is throttled back in presence of other active streams.
    • The "stripe pool" resource is automatically shared evenly between all active streams.

     

    As a result got rid of some Tunables:

    • md_sync_window
    • md_sync_thresh

    and added some tunables:

    • md_queue_limit
    • md_sync_limit
    • [-rc2] md_scheduler

     

    Please refer to Settings/Disk Settings help text for description of these settings.


    WireGuard support - available as a plugin via Community Apps.  Our WireGuard implementation and UI is still a work-in-process; for this reason we have made this available as a plugin, though the latest WireGuard module is included in our Linux kernel.  Full WireGuard implementation will be merged into Unraid OS itself in a future release.  I want to give special thanks to @bonienl who wrote the plugin with lots of guidance from @ljm42 - thank you!  I also should give a shout out to @NAS who got us rolling on this.  If you don't know about WireGuard it's something to look into!

     

    Guide here:


    WS-Discovery support - Finally you can get rid of SMBv1 and get reliable Windows network discovery.  This feature is configured on the Settings/SMB Settings page and enabled by default.

    • Also on same settings page is Enable NetBIOS setting.  This is enabled by default, however if you no longer have need for NetBIOS discovery you can turn it off.  When turned off, Samba is configured to accept only SMBv2 protocol and higher.
    • Added mDNS client support in Unraid OS.  This means, for example, from an Unraid OS terminal session to ping another Unraid OS server on your network you can use (e.g., 'tower'):
      ping tower.local
      instead of
      ping tower
      Note the latter will still work if you have NetBIOS enabled.

     

    User Share File System (shfs) changes:

    • Integrated FUSE-3 - This should increase performance of User Share File System.
    • Fixed bug with hard link support.  Previously a 'stat' on two directory entries referring to same file would return different i-node numbers, thus making it look like two independent files.  This has been fixed however there is a config setting on Settings/Global Share Settings called "Tunable (support hard links)".  [-rc2 ] Fixed the default value Yes, but with certain very old media and DVD players which access shares via NFS, you may need to set this to No.
      [-rc5] Fixed not accounting for devices not mounted yet.
    • Note: if you have custom config/extra.cfg file, get rid of it.


    Other improvements/bug fixes:

    • Format - during Format any running parity sync/check is automatically Paused and then resumed upon Format completion.
    • Encryption - an entered passphrase is not saved to any file.
    • Fixed bug where multi-device btrfs pool was leaving metadata set to dup instead of raid1.
    • Several other small bug fixes and improvements.
    • [-rc5] Fixed bug where quotes were not handled properly in passwords.
    • Numerous base package updates [-rc2] including updating PHP to version 7.3.x, Samba to version 4.11.x.

     

    Known Issues and Other Errata

    • Some users have reported slower parity sync/check rates for very wide arrays (20+ devices) vs. 6.7 and earlier releases - we are still studying this problem.
    • [-rc6] this is fixed: If you are using Unassigned Devices plugin with encrypted volumes, you must use the file method of specifying the encryption passphrase.  Note that a file containing your passphrase must consist of a single null-terminated string with no other line ending characters such as LF or CR/LF.
    • In another step toward better security, the USB flash boot device is configured so that programs and scripts residing there cannot be directly executed (this is because the 'x' bit is set now only for directories).  Commands placed in the 'go' file still execute because during startup, that file is copied to /tmp first and then executed from there.  If you have created custom scripts you may need to take a similar approach.
    • AFP is now deprecated and we plan to remove support in Unraid 6.9 release.

     

    A note on password strings

    Password strings can contain any character however white space (space and tab characters) is handled specially:

    • all leading and trailing white space is discarded
    • multiple embedded white space is collapsed to a single space character.

    By contrast, encryption passphrase is used exactly as-is.

     

    Version 6.8.0-rc9 2019-12-06

    Base distro:

    • btrfs-progs: version 5.4
    • ca-certificates: version 20191130
    • ebtables: version 2.0.11
    • gnutls: version 3.6.11.1
    • gtk+3: version 3.24.13
    • iproute2: version 5.4.0
    • iptables: version 1.8.4
    • keyutils: version 1.6
    • libepoxy: version 1.5.4
    • libnftnl: version 1.1.5
    • librsvg: version 2.46.4
    • libtasn1: version 4.15.0
    • lvm2: version 2.03.07
    • nano: version 4.6
    • pcre2: version 10.34
    • pkgtools: version 15.0 build 28
    • wireguard: version 0.0.20191205
    • xorg-server: version 1.20.6

    Linux kernel:

    • version 4.19.88
    • Like 3
    • Thanks 6



    User Feedback

    Recommended Comments



    19 minutes ago, glennv said:

    @Tucubanito07 Nope , you can like i have use a large SSD based cache and set your share(s) to cache=yes/only.

    Then after the initial memory cache the writes go to the fast ssd's (or nvme's Like Johnnie has) .

    You can then at a convenient time uffload to spindels (mover) or keep in in the cache, whatever suits your workflow

    Thank you. I will test on this today when I get home. 

    Link to comment

    I'm seeing an issue since rc8 and now rc9.  I'm getting random read errors on all my hard drives.  System is freshly rebuilt with:

    6 drives (4 Seagate 14TB, 2 Seagate 16TB)

    ASRock Z390 motherboard with I9-9900 CPU

    128GB Ram

    1TB NVME for cache

    Drives are spread across onboard controller and two Broadcom SAS3008 controllers

     

    Read errors start showing up within a few hours of upgrading.  When I revert back to rc7 and run a parity check, no errors are found.  The server has several docker containers running, but no VM's at the moment.

     

    I started running the RCx series for hardware transcoding, and it has worked perfectly prior to rc8.

     

    Let me know if I can add any additional information.

     

                 Dan

    Link to comment

    Sadly the performance issues I'm having with 6.8.x continue.

     

    Downloading torrents pretty much grinds to a halt compared to 6.7.2.  There's something different in the sauce between these two that's just not working properly.  File transfers seem OK, but torrent chunks coming in seem to be very problematic.  Get an intial burst that comes in quickly, then everything else just stalls out afterwards and get periodic small transfer bursts as others have been in here.

     

    A straight file copy doesn't seem to be a problem and can truck along at ~112 MB/Sec.  But a torrent download spikes to ~10MB/sec and then just trickles after that.

     

    Revert back to 6.7.2 and golden again.  Making me cry, 'cause I really like the new login page. Diagnostic attached.

    kizunaai-diagnostics-20191210-2022.zip

    Link to comment
    15 hours ago, dan-dekalb said:

    Read errors start showing up within a few hours of upgrading.  When I revert back to rc7 and run a parity check, no errors are found.  The server has several docker containers running, but no VM's at the moment.

    rc7 is on 5.x kernel, rc8/9 is on 4.x kernel.

    Could be your problem if rc7 is fine and rc8/9 is not.

     

    If your hardware works well under rc7 then you are better off waiting for 6.9.0-rc1 (which LT has publicly announced that it will come out shortly after 6.8.0 stable is out). It is the same as 6.8.0 but with 5.x kernel.

     

    I'm holding out for that too.

    Link to comment

    @joelones I found a solution for the Pfsense issue. Looks like the cpu-mode "host-passthrough" is the culprit. Either switch to "Emulated QEMU64" or emulate a Skylake CPU for example.

    Not sure what changed in the current Qemu version which caused this issue. This might be also the reason why a couple people are reporting problems with their VMs won't boot up after an update of Unraid. Maybe something @limetech should know and have a deeper look into.

     

    I'am currently on RC9 and don't know when this issue started. As mentioned earlier, I fire this VM up only when I need it for testing and can't really remember on which Unraid build I used it the last time. Never had an issue with 'host-passthrough' before.

    Link to comment

    @limetech I didn't want to open a new thread for this as its tiny issue, when trying to change the description of my server with the new update I cant use "&" the "&" and anything I type after will disappear.

    Link to comment
    On 12/11/2019 at 5:35 AM, dgriff said:

    Sadly the performance issues I'm having with 6.8.x continue.

     

    Downloading torrents pretty much grinds to a halt compared to 6.7.2.  There's something different in the sauce between these two that's just not working properly.  File transfers seem OK, but torrent chunks coming in seem to be very problematic.  Get an intial burst that comes in quickly, then everything else just stalls out afterwards and get periodic small transfer bursts as others have been in here.

     

    A straight file copy doesn't seem to be a problem and can truck along at ~112 MB/Sec.  But a torrent download spikes to ~10MB/sec and then just trickles after that.

     

    Revert back to 6.7.2 and golden again.  Making me cry, 'cause I really like the new login page. Diagnostic attached.

    kizunaai-diagnostics-20191210-2022.zip 142.31 kB · 2 downloads

    Im seeing the same exact thing - I started out on 6.7.2 but didnt test it a lot but then moved to the RC versions, mainly to get BBR - But now that my torrent archive is starting to grow again, after the migration im getting hit this more and more, I will try and downgrade to 6.7.2 to see if its any better for me there and then just hope that I wont hit the corruption issue as I would hate that.

    *Edit - can confirm that downgrading to 6.7.2 resolves this, performance on torrent workloads are a lot faster on 6.7.2 compared to 6.8, im seeing the same complete stalls when there's writes going on, on torrents.

    Edited by kallimeo
    Link to comment
    On 12/13/2019 at 1:37 PM, kallimeo said:

    Im seeing the same exact thing - I started out on 6.7.2 but didnt test it a lot but then moved to the RC versions, mainly to get BBR - But now that my torrent archive is starting to grow again, after the migration im getting hit this more and more, I will try and downgrade to 6.7.2 to see if its any better for me there and then just hope that I wont hit the corruption issue as I would hate that.

    *Edit - can confirm that downgrading to 6.7.2 resolves this, performance on torrent workloads are a lot faster on 6.7.2 compared to 6.8, im seeing the same complete stalls when there's writes going on, on torrents.

    I'm not sure if it's possible to get a special build of 6.8 without the mitigations included to see if that solves the problem, but I'd be willing to help try and track this down as it's basically unusable for me as an upgrade right now.  I didn't have the corruption issue before anyway as everything with databases are on my cache drives.

    Link to comment
    6 hours ago, dgriff said:

    I'm not sure if it's possible to get a special build of 6.8 without the mitigations included to see if that solves the problem, but I'd be willing to help try and track this down as it's basically unusable for me as an upgrade right now.  I didn't have the corruption issue before anyway as everything with databases are on my cache drives.

    There is a      Disable Mitigation Settings      plugin available through the Apps plugin...

    Link to comment
    3 hours ago, Frank1940 said:

    There is a      Disable Mitigation Settings      plugin available through the Apps plugin...

    I think this more relates to the mitigations on the md-driver to try and fix the issue with the mover in 6.7.2 killing all reads.

     

    I dont have any issues either with corruption as all my databases are on UD drives, but performance is much faster on 6.7.2 for torrents, I just did a test now and on 6.8 a torrent downloaded 8.1GB with avg of 4.3MB\s - Dropping back to 6.7.2 got that same torrent down with an avg of 191MB\s (I have 10G internet), so a massive difference.

    Link to comment
    2 minutes ago, kallimeo said:

    I dont have any issues either with corruption as all my databases are on UD drives, but performance is much faster on 6.7.2 for torrents, I just did a test now and on 6.8 a torrent downloaded 8.1GB with avg of 4.3MB\s - Dropping back to 6.7.2 got that same torrent down with an avg of 191MB\s (I have 10G internet), so a massive difference.

    Is that downloading to cache or array? Changes to md driver would only affect if downloading directly to the array.

    • Thanks 1
    Link to comment
    On 12/18/2019 at 3:30 PM, johnnie.black said:

    Is that downloading to cache or array? Changes to md driver would only affect if downloading directly to the array.

    Sorry about the late reply.

     

    All of my torrents run to my array, I only have one share which is set to use the cache drive - Performance is "fine" on 6.7.2 - Not super great, but it is okay, still seeing my rtorrent instance being heavily throttled by the fuse system way before the drives are really loaded in any way, but I can live with that - On 6.7.2 the same exact setup runs fine and I get files down at basically ~200MB\s, but on 6.8 it will burst really fast starting out, then it will just drop toe 3-5MB\s avg

    unraid.png

    Link to comment
    2 minutes ago, kallimeo said:

    All of my torrents run to my array

    You should open a bug report and please post the diagnostics, ideally grabbed during a slowdown.

    Link to comment
    1 minute ago, johnnie.black said:

    You should open a bug report and please post the diagnostics, ideally grabbed during a slowdown.

    Thanks - I will do that in the new year, im fine by running 6.7.2 as I dont really have the slowdowns on that and the sqlite issue isnt there either since I use UD devices for those.

     

    I do think that read performance for torrent workloads are better on 6.8 - I generally pushed more MB per day on that vs 6.7.2 so something has improved.

    Link to comment

    Is it possible to revert back to basic authentication? It's useful for externally grabbing data for example for my JSON API 

     

    Perhaps a option? I haven't personally checked yet but had a issue raised on my plugin about it not working on 6.8

     

    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
    Add a comment...

    ×   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.


  • Status Definitions

     

    Open = Under consideration.

     

    Solved = The issue has been resolved.

     

    Solved version = The issue has been resolved in the indicated release version.

     

    Closed = Feedback or opinion better posted on our forum for discussion. Also for reports we cannot reproduce or need more information. In this case just add a comment and we will review it again.

     

    Retest = Please retest in latest release.


    Priority Definitions

     

    Minor = Something not working correctly.

     

    Urgent = Server crash, data loss, or other showstopper.

     

    Annoyance = Doesn't affect functionality but should be fixed.

     

    Other = Announcement or other non-issue.