Leaderboard

CA Config Editor A simple plugin to allow you to easily edit any of unRaid's configuration files (or actually any file on your server) within your browser without having to go to the command line. (including syslinux.cfg) While most of unRaid's settings are available to be modified within its webUI, there are some cases where it is required / desired to manually edit them. This plugin will allow you to do that. Backups of files edited are automatically created (filename.bak), and can be restored by copying from the backup window. Additionally, any file present on your server in any folder can also be edited beyond the .cfg files that shortcuts are available for. This is helpful for when a docker application requires you to make some changes within its config files in the appdata folder, but either permissions prevent you from doing this over the network, or you're not familiar with the command line. The usual warnings apply: Changes to the configuration files could have unintended results if you enter in wrong values. Install it via the Apps Tab (where else?). Either go to the CA Modules section, or search for Config and scroll down to CA Config Editor. Once installed, you can find the plugin within the Tools tab.

Apache Guacamole is a clientless remote desktop gateway. It supports standard protocols like VNC, RDP, SSH, and Telnet. This docker primarily has a MariaDB (MySQL) database built-in for authentication and configuration. It also has support for LDAP authentication and configuration as well as Duo two-factor authentication. For general usage of Apache Guacamole the full manual is located here: https://guacamole.apache.org/doc/gug/ The different authentication methods can be enabled or disabled through the template. MySQL is the default authentication and configuration method. Template Options OPT_MYSQL: Y(default) / N OPT_MYSQL_EXTENSION: Y / N(default) OPT_SQLSERVER: Y / N(default) OPT_LDAP: Y / N(default) OPT_DUO: Y / N(default) OPT_CAS: Y / N(default) OPT_OPENID: Y / N(default) OPT_TOTP: Y / N(default) OPT_QUICKCONNECT: Y / N(default) Note: MySQL, SQL Server, and LDAP cannot be used at the same time. Duo two-factor can be used with MySQL, SQL Server, or LDAP authentication. Microsoft SQL Server Authentication/Configuration For SQL Server support modify your guacamole.properties file as follows: Create a database on your SQL Server Execute the SQL scripts found in the sqlserver-schema folder on the new database you created. Comment the MySQL properties. Uncomment the SQL Server properties. Modify SQL Server properties to match your configuration. Further instructions can be found here: https://guacamole.apache.org/doc/gug/jdbc-auth.html#jdbc-auth-sqlserver LDAP Authentication/Configuration For LDAP authentication modify your guacamole.properties file as follows: Comment the MySQL properties. Uncomment the LDAP properties. Modify LDAP properties to match your configuration. Further instructions can be found here: https://guacamole.apache.org/doc/gug/ldap-auth.html Duo Two-Factor Authentication For Duo two-factor authentication uncomment the Duo properties in your guacamole.properties and follow these instructions: https://guacamole.apache.org/doc/gug/duo-auth.html#guac-duo-config 4/4/2017 Upgraded to Apache Guacamole 0.9.12-incubating. Upgraded MySQL Java Connector to 5.1.41. 8/2/2017 Upgraded to Apache Guacamole 0.9.13-incubating. Upgraded MySQL Java Connector to 5.1.43. Database schema upgrades to 0.9.13. 8/4/2017 Added all optional libraries for guacd to enable all features. Got rid of log warning by moving the offending symbolic link creation from firstrun.sh to Dockerfile. 8/7/2017 Upgraded to phusion/baseimage 0.9.22. Upgraded to MariaDB 10.2. Upgraded to Tomcat 8. 12/16/2017 Added Powerline fonts to support Oh My Zsh agnoster theme. 1/24/2018 Upgraded to Apache Guacamole 0.9.14. Upgraded MySQL Java Connector to 5.1.45 Added Microsoft SQL Server JDBC 6.2.2 driver. Converted to multi-stage Docker to reduce size. Ability to downgrade using a tag (e.g. jasonbean/guacamole:0.9.13) 6/7/2018 Added GhostScript to allow remote printing. 6/8/2018 Changed location of Tomcat catalina.out log file /config/guacamole/log to improve troubleshooting. Attempt to mitigate a file permissions problem by always setting 755 permissions on container start. 6/11/2018 Fixed creation of /config/guacamole/log folder so that it will exist when Tomcat is started. Corrected a regression error introduced when switching to a multi-stage build. 6/12/2018 Modified unRAID template to make this a privileged container so that Tomcat will run correctly. Added more logging messages and ensure that the database doesn't get created when the guacamole.properties file doesn't exist. Modified so that all Tomcat logs get placed in guacamole/log directory. 8/18/2018 tyler43636 submitted pull request to add the tzdata package to correct for the RTC being set to the local time rather than UTC. 9/24/2018 Modified the firstrun.sh script so that the MySQL schema files are copied to /config when OPT_MYSQL is set "Y" in case someone wants to use an external MySQL or MariaDB database. 1/10/2019 Upgraded to Apache Guacamole 1.0.0. BIG UPGRADE! Thanks to the Guacamole team my build process has been greatly improved. Now uses the Debian base container just like the guacd Docker container from Apache. Retooled entire run process to use tini/supervisor to manage services. Added version that comes without MariaDB so that container size is halved if you have an external database source or use the default authentication. 2/23/2020 Upgraded to Apache Guacamole 1.1.0. Changed to debian:stable-slim base container to shave off some size. No database schema changes. 7/5/2020 Upgraded to Apache Guacamole 1.2.0. Fixed problem with OpenID extension so that it would load before all other auth extensions. No database schema changes. 11/29/2020 Added -s argument to tini to make sure it can subreap if not PID 1. 12/6/2020 Implemented PUID and PGID environment variables to run tasks in non-root and match nobody user in file system. Cleaned up code to improve maintainability. 12/14/2020 Fixed regression to restore OPT_MYSQL_EXTENSION. 1/9/2021 Upgraded to Apache Guacamole 1.3.0. 5/30/2021 Merged changes from Sean to add logrotate to prevent Catalina log from growing forever. 1/12/2022 Upgraded to Apache Guacamole 1.4.0. Changed to install latest version of Tomcat directly from Apache since the Debian Buster repository contains an old version. 11/11/2022 Updated MariaDB install process. Cleaned up code. New build to update Tomcat. 3/18/2023 Correct how mysqld loads so that it shuts down correctly. 3/21/2023 Upgraded to Apache Guacamole 1.5.0. 3/24/2023 Added missing dependency for guacd. 4/14/2023 Upgraded to Apache Guacamole 1.5.1. Removed privileged from template. Changed to OpenJDK 11 to fix OpenID. 5/26/2023 Upgraded to Apache Guacamole 1.5.2. Implmented client log level. Simplified Dockerfile. 8/4/2023 Upgraded to Apache Guacamole 1.5.3. 1/10/2023 Upgrade to Apache Guacamole 1.5.4. Simplify versioning.

Edit: working on full guide now as I've cracked it Happy to share when I've got time to put something together that hopefully someone else can follow and understand what I did! The traffic shaper in pfSense needs work I think, although I think the fact that my WAN traffic is all going via AirVPN made it harder to setup. I'll run through what I finally settled upon (or stumbled upon....) last night as an acceptable solution after literally dozens of attempts over several days to shape my traffic, and I'll do a proper guide later. In brief, pfSense offers three ways of shaping. The most basic "PRIQ" assigns a priority of 0-7 with seven being the highest where traffic with priority 7 gets bandwidth first and priority 6 doesn't until 7 has taken all it wants, then priority 5 and so on. The problem here is that certain services can hog all the bandwidth rather than ensuring everyone gets a 'little' e.g. setting POP3 at priority 3 might mean that it NEVER gets any bandwidth i.e. emails would never get sent if 4-7 are using up all the bandwidth. The most advanced method "HFSC" fixes this problem by creating queues and allowing you to set that even if the WAN or LAN is maxed out, that each of the queues gets a min amount of bandwidth (can be set as a % of the parent queue or an absolute amount e.g. 1Mb) e.g. the default queue for qP2P is given 5% of a backlogged queue, qVoip is given a guaranteed configurable min to ensure call quality remains acceptable. My solution (will do in more detail in a proper post): ran the wizard and choose dedicated links as I have 1 LAN and 1 WAN connection set my line speeds based on my broadband connection - 18Mbps and 1Mbps Chose the types of traffic I wanted to prioritise - VoiP, P2P, NNTP, SMTP, Hangouts, individual games etc etc Ended up with default rules for LAN (won't talk about WAN and qACK for now - just want to show potential and how easy it was for me to eventually create a solution that worked for me): LAN - PARENT initially set at 18Mbps max via wizard ---qLink - child of LAN. I think this queue is for internal LAN traffic and personally is why I think the shaper is so hard to figure out. More below ---qInternet - child of LAN. pfSense gave this 80% of LAN bandwidth, and created the children below: --------qACK --------qP2P - set at 5% of parent qInternet bandwidth if backlogged --------qVoiP --------qGames - set at 20% I think of parent qInternet bandwidth if backlogged. Selected games added here by pfSense through automatic rules --------qOthersHigh - set at 20% of parent qInternet bandwidth if backlogged. Selected protocols added here that I asked to prioritise in the wizard e.g. Hangouts --------qOthersLow - set at 5% of parent qInternet bandwidth if backlogged. i added things like email here I then created another queue qSabnzbd under Parent qInternet, and created a floating rule to put all traffic on port 563 here (pfSense only created a rule for port 119) - did this so I could see if it worked by watching the impact in Status/Queues Now the problem I had, was that some of my non-unRAID internet traffic was getting thrown into qLink (think the VMs were the problem here as I don't think pfSense could work out where the traffic was coming from and going to properly), so impossible to shape (or I couldn't work out how to) - I can't limit qLink because it would start impacting internal traffic e.g. streaming a film to a TV.... P2P traffic was a big culprit. I've double-checked and my internal LAN traffic isn't going into qLink, just my P2P and some other random internet traffic that is trickling through. I solved this by setting the 'bandwidth share of a backlogged queue' for P2P at 20%, so that if my line is maxed out from qInternet, then P2P via qLink would be capped at 20% of bandwidth. I could set this lower, but until I'm certain what other internet traffic is going into qLink, I don't want to starve this queue. I fixed this by: I then made the following tweaks: Increasing LAN to 1Gbps to match my actual network Setting child qLink at 986Mbps Setting child qInternet at 14Mbps scheduling my qBittorrent client to not exceed 2Mbps during the day and unlimited overnight Why this works for me? Setting qLink at 20% max share 986Mbps means internal traffic internet traffic that's allocated to qLink always gets through as planned by pfSense without shaping During the day, my P2P traffic that's unfortunately in qLink is locked at a max of 2Mbps - if I didn't it would take all my bandwidth. Any other internet traffic that leaks through to qLink will get priority - again ok with me as this will be user-initiated and not background traffic like P2P or Sabnzbd. I've assumed that this will be at most a couple of Mbps I set qInternet at 14Mbps to ensure that the bulk of my internet traffic gets shaped e.g. if I'm watching a youtube video (in the qOthersHigh), then it will take bandwidth away from qSabnzbd (qSabnzbd queue is allocated only 5% of qInternet in a backlogged queue). qOthersHigh is allocated a guaranteed 20% of a backlogged queue i.e. 3.6Mbps, so if the youtube video needs 3Mbps then qSabnzbd (and all over traffic) will use a maximum of 15.4Mbps 11Mbps, and jump back up to 184Mbps when the video is finished overnight when we're asleep it's ok for P2P via qLink to nab all the traffic with qInternet taking whatever's left It's working brilliantly, because as soon as we start browsing or want to watch a video or stream something, pfSense immediately takes traffic away from lower priority tasks e.g. qSabnzbd and gives it back as soon as we've finished. It means my pitiful 18Mbps (they are digging fibre in my village so not for much longer) is working so much more efficiently, prioritising foreground traffic and giving it to background traffic immediately when not needed. The only drawback is if more than 4Mbps of internet traffic leaks into qLink then it will mean that qInternet won't get shaped because less than 14Mbps will be going through that queue, so pfSense won't shape. This should hopefully be rare, but hopefully in the future I'll work out how to get all my internet traffic through the qInternet queue. That was a lot longer than I planned for a brief teaser!

Sigs don't show in Tapatalk or mobile browsers (at least on iPhone). You have to use a desktop browser to see them.

The best way is to install the 'Fix Common Problems' plugin and then put it into troubleshooting mode.

I had no idea my Spanish was so good

Thanks to my friend Vern over in Calgary for emailing me this good news this morning, I have learnt finally the wait is almost over and Nvidia is releasing beta osx web drivers next week for osx! So, at last, i can use my 10 series card in osx. News was announced with the release of the Titan XP https://9to5mac.com/2017/04/06/nvidia-titan-xp-beta-pascal-drivers-mac/

That's a lot of errors on the Samsung SSD, and almost certainly the root cause of the docker issues. That SSD is on a Marvell controller, that would be my prime suspect, and I would do the following: -shutdown the server and swap both power and SATA cables with the Corsair SSD -power back up and type: btrfs device stats -z /mnt/cache This will show current stats once again but reset all values to 0 -run a scrub, if any errors are found correct them -continue using the server normally and every day or every couple of days check stats again with: btrfs device stats /mnt/cache If more errors appear and before rebooting grab and post new diagnostics, if no more errors appear for a few weeks you're probably fine.

If your only problem with the SAS2LP is slow parity checks that can be fixed with the right tunnables, not worth replacing them unless you're having other issues, like sync errors or disks dropping out.

So to recap: The UD plugin will display a "Smart Attributes" link (which shows the Self-Test, Attributes, Capabilities, and Identity tabs) for an unassigned device if: a. The drive is mounted b. -or- if you have defined a device script If the drive is not mounted and there is no script, UD assumes it is a warm spare and will not do anything that might spin the drive up. It sounds like stock unRAID will display this link at all times, although I have not tested it. -- Here are some potential enhancement requests based on this discussion: Provide some way to enable SMART on unassigned devices, if the device supports it but it is disabled by default. Otherwise the user has to drop to the command line to enable SMART. If this functionality is provided by unRAID it can either be fully automatic or it could be an option on the Identity tab (i.e. provide an "Enable" button next to the line that says SMART support is disabled) This could potentially be automated by the UD plugin Or it could be handled by @gfjardim's preclear plugin, by having it automatically enable SMART before a preclear rather than just telling you it is disabled. This options seems the least intrusive. Modify unRAID to allow custom SMART Settings and temperature thresholds to be set for unassigned devices, and modify the back end scripts to use them. Currently these settings are only available to array drives. Consider modifying UD to make it more clear why the "Smart Attributes" link is disabled. This seems to be a source of confusion.

This is pretty much the only thing missing. In this day and age, and being a victim of a home invasion, we really need something even if it is just encfs. For now, I back up important data to the cloud/friend's server using rclone. But still doesn't solve the issue of an outright server being stolen. Though, the 'puter is so heavy... I hope nobody would bother picking it up, and rather go for Tablets, and Laptops. And ofcourse, they love taking Gift cards. Moral: Don't keep physical gift cards. Apparently, they're not that interested in stealing credit cards either. Though YMMV. Sorry, didn't mean to be 'preachy.'

I recommend using 1 SSD for cache and avoiding BTRFS entirely. It's not mature enough, yet.

It seems to be quite normal that when installing Windows 10 that you need to limit the VM to using a single core. After installation has completed you can increase the number. Why this I have no idea but it seems to be a common issue.

4x 3TB WD Red drives, one for parity, room for two more drives 2x 250GB Samsung SSDs for cache pool Little less than half of the available space is in use. Backed up my own DVDs, use Plex Docker to stream to my XBox/TV. Don't download media unless purchased. Trying out some Docker'd applications for utilities, backups etc. Already using a backup share to backup my main PC.

I have seen the same on a Fedora 25 VM