Drabert Posted April 17, 2020 Share Posted April 17, 2020 UnRaid 6.8.3 with Nvidia GPU enabled Hey Everyone... I was trying to add some new disks that were not formatting correctly so i added the "unassigned Devices" plugin and "binhex-preclear" to my box... something got horked and now my UI looks like gunk and this message is showing: Warning: parse_ini_file(/boot/config/plugins/dynamix/dynamix.cfg): failed to open stream: Success in /usr/local/emhttp/plugins/dynamix/include/Wrappers.php on line 22 Warning: array_replace_recursive(): Expected parameter 2 to be an array, bool given in /usr/local/emhttp/plugins/dynamix/include/Wrappers.php on line 22 Warning: extract() expects parameter 1 to be array, null given in /usr/local/emhttp/plugins/dynamix/include/DeviceList.php on line 24 it got funky as soon as i added the binhex-preclear plugin so i stopped the plugin and removed it, but still no love. I uninstalled the dynamix plugin and rebooted, reinstalled the dynamix plugin, rebooted again, still getting the wierd messages. Everything else is working fine though, all of the docker containers seem to function fine. Is there an easy way to just update UnRAID to get the system back to its "normal" state? Thanks Drabert Quote Link to comment
bonienl Posted April 17, 2020 Share Posted April 17, 2020 Start in safe mode and recheck. Quote Link to comment
Drabert Posted April 17, 2020 Author Share Posted April 17, 2020 Will do... thank you for the advice Quote Link to comment
Drabert Posted April 18, 2020 Author Share Posted April 18, 2020 I tried rebooting in safe mode but got the same error. any other suggestions? Quote Link to comment
JorgeB Posted April 18, 2020 Share Posted April 18, 2020 Backup flash drive, recreate a new flash with the USB tool, restore only super.dat and the key (both on the config folder), it should now boot correctly like that, than either restore the other settings a few at time until you find the problem or just reconfigure the server. Quote Link to comment
Drabert Posted April 18, 2020 Author Share Posted April 18, 2020 will that keep all of my docker vms intact or will I have to rebuild them? Quote Link to comment
JorgeB Posted April 18, 2020 Share Posted April 18, 2020 It will keep them all, before re-enabling them just make sure both services are using the previous existing paths, settings -> docker and settings -> vm manager Quote Link to comment
Squid Posted April 18, 2020 Share Posted April 18, 2020 Also, grab the contents of /config/plugins/dockerMan/templates-user from the original flash. Quote Link to comment
kenyonj Posted April 25, 2020 Share Posted April 25, 2020 I also experienced this immediately after installing `binhex-preclear`. Was there something short of replacing the flash contents that fixed this for you @Drabert? Quote Link to comment
Squid Posted April 25, 2020 Share Posted April 25, 2020 19 minutes ago, kenyonj said: I also experienced this immediately after installing `binhex-preclear`. It has nothing to do with any given docker container being installed. Quote Link to comment
Drabert Posted April 25, 2020 Author Share Posted April 25, 2020 Sorry. I have not had a chance to fix it yet. The system is still running fine, the GUI is just fugly Hoping to try to do what you all recommended this evening to see if that resolves my issue... it just took a while to get it all working, and I am afraid that something during the reinstall will nuke the VMs i have already created. I am running the NVIDIA branch of 6.8.3... was thinking of just reverting back to the stock 6.8.3 to see if that rebuilds whatever is stuck and then redo NVIDIA. Would that work too? Quote Link to comment
Drabert Posted April 26, 2020 Author Share Posted April 26, 2020 So i went ahead and reinstalled unRaid, added the License key and super.dat to the new image. The UI is back to normal, but now i lost the ability to "edit" all of my docker components and a couple of them are failing to start. I went back and also added the files in /config/plugins/dockerMan/templates-user/ and i was able to edit the components again. I am installing the NVIDIA plugin right now... thank you for the help! Quote Link to comment
M3K Posted April 28, 2020 Share Posted April 28, 2020 On 4/25/2020 at 9:59 AM, Squid said: It has nothing to do with any given docker container being installed. I also experienced this immediately after installing binhex-preclear. Quote Link to comment
Sh0ttas Posted May 2, 2020 Share Posted May 2, 2020 Same here with the binhex-preclear. Quote Link to comment
Squid Posted May 2, 2020 Share Posted May 2, 2020 On 4/25/2020 at 9:59 AM, Squid said: It has nothing to do with any given docker container being installed. Quote Link to comment
imsplitbit Posted May 7, 2020 Share Posted May 7, 2020 (edited) This empirically is due to the binhex-preclear install but only indirectly. When you install that plugin it asks you for volume mounts to give to the underlying docker daemon for mounting inside the docker container. Host Path 6 is /boot/config/plugins/dynamix/dynamix.cfg. This *should* be a file, I'm assuming, but didn't exist on my unraid. When you give docker a path that doesn't exist it will automatically create a folder given that path. So when I installed binhex-preclear the docker daemon created the folder /boot/config/plugins/dynamix/dynamix.cfg/ If you make this folder and refresh your WebUI you will see that it generates php errors described above. The *easiest* fix is: 1. Uninstall the binhex-preclear plugin/app 2. remove that path from your dynamix plugins folder (or at the very least move it to another name that *doesn't* end in .cfg) * You can move the offending directory out of the way via the terminal with the command: * mv /boot/config/plugins/dynamix/dynamix.cfg /boot/config/plugins/dynamix/dynamix.cfg.broken * Or if you want to remove it * rmdir /boot/config/plugins/dynamix/dynamix.cfg 3. Now refresh the WebUI and you'll see the errors are gone. Now if you want to reinstall the binhex-preclear plugin you should first create that file that the plugin wants to pass through to the docker container. From the terminal run the command: touch /boot/config/plugins/dynamix/dynamix.cfg [EDIT] Squid pointed out that if you make changes within Settings -> Display Settings this will create the file /boot/config/plugins/dynamix/dynamix.cfg so there's no need to enter the terminal if you don't want to. Then proceed with the install as you normally would and things should operate just fine. Edited May 7, 2020 by imsplitbit 3 Quote Link to comment
Squid Posted May 7, 2020 Share Posted May 7, 2020 2 hours ago, imsplitbit said: but didn't exist on my unraid. Which means that you've never made any changes within Settings - Display Settings. @binhex Might want to add this to your FAQ or something. Quote Link to comment
imsplitbit Posted May 7, 2020 Share Posted May 7, 2020 (edited) That's probably true. I'm very new to unraid but have been a linux systems engineer for 20 years so I'm happy to help folks when/where I can. This array is only about 12 hours old but it did have issues immediately following the install of binhex-preclear. After reading this thread I was about to do a backup and restore as detailed above but thought I'd spend a little time digging into the problem to see if all of that was necessary. Edited May 7, 2020 by imsplitbit 1 Quote Link to comment
bonienl Posted May 7, 2020 Share Posted May 7, 2020 Your investigations may certainly help others. Thanks Quote Link to comment
binhex Posted May 7, 2020 Share Posted May 7, 2020 (edited) thanks for the detailed debug @imsplitbit, i was not aware that the dynamix.cfg could not exist, i assumed some default values would always be required and thus the file would be generated, my bad. ok simple fix for this is to switch the volume mount point so that it no longer points at a specific file but the folder instead, example modification (unraid docker template):- <Volume> <HostDir>/boot/config/plugins/dynamix/</HostDir> <ContainerDir>/unraid/config/plugins/dynamix/</ContainerDir> <Mode>ro</Mode> </Volume> the effect is the same, it just exposes all files in that folder as opposed to just one, but as the volume is read only it shouldnt mean the risk of users accidentally messing with the host files is any greater so should be fine, sound good?. Edited May 7, 2020 by binhex Quote Link to comment
binhex Posted May 11, 2020 Share Posted May 11, 2020 for anybody following this thread, my change is now in and hostdir and containerdir are now both pointing at path rather than directly at file dynamix.cfg, i have checks already in the code to check if the file exists so that will be covered in the case where the file doesnt exist (as is the case with somebody who hasnt configured anything in unraid web ui settings). 1 Quote Link to comment
betaman Posted December 14, 2020 Share Posted December 14, 2020 (edited) Just read thru last couple pages of this thread. Seems like everything is taken care of (known bugs anyway) but was looking for some confirmation before clearing a 10TB drive with this docker. Also, my parity is currently rebuilding and the new disk is unassigned and not formatted. I guess I need to wait to stop array and format or does the preclear script also do this per my disk setting (ie MBR 4K aligned)? I think I found my answer In FAQ: Alignment Options (-a and -A flags) Below is the logic built into the preclear script when determining the starting sector for the partition:- if preclear -a option chosen and drive < 2.2TB then will set to 63. if preclear -a option chosen and drive > 2.2TB then will set to 64. if preclear -A option chosen and drive < 2.2TB then will set to 64. if preclear -A option chosen and drive > 2.2TB then will set to 64. if preclear -a or -A not defined then 63 or 64 will be chosen based on unraid config, UNLESS the drive is > 2.2TB, in which case it will always be 64. Edited December 14, 2020 by betaman Quote Link to comment
itimpi Posted December 14, 2020 Share Posted December 14, 2020 8 minutes ago, betaman said: I guess I need to wait to stop array and format or does the preclear script also do this per my disk setting (ie MBR 4K aligned)? You must add a disk to Unraid BEFORE trying to format it as formatting a disk destroys its ‘Clear’ status. Quote Link to comment
betaman Posted December 14, 2020 Share Posted December 14, 2020 (edited) 5 minutes ago, itimpi said: You must add a disk to Unraid BEFORE trying to format it as formatting a disk destroys its ‘Clear’ status. Disk is “added” but unassigned. I want to preclear the disk so it’s ready in case of a drive failure. I just don’t remember seeing the format notice on a device before but I haven’t added or precleared a drive in quite some time! So the order of operations is preclear, format, assign? Edited December 14, 2020 by betaman Quote Link to comment
trurl Posted December 14, 2020 Share Posted December 14, 2020 4 minutes ago, betaman said: Disk is “added” but unassigned. Added means ADDED to the array. 5 minutes ago, betaman said: So the order of operations is preclear, format, assign? So no this is not correct. If you format the precleared disk before you assign it, then it won't be clear anymore and Unraid will have to clear it. 10 minutes ago, itimpi said: You must add a disk to Unraid BEFORE trying to format it as formatting a disk destroys its ‘Clear’ status. Quote Link to comment
Recommended Posts
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.