meep

So with Docker and VMs disabled, the system is still generating multiple GPFs and Kernel crashes. I've attached todays Syslog that shows a boot up sequence around 8:40, with GPFs starting after 10:00. I did manage to do a clea shutdown, so there's that. Next, I'm going to remove any additional non-essential hardware such as extra GPUs, USB conbtroller etc. If still problematic, I'll boot to safe mode to eliminate PlugIns and after that, it's going to be a CPU and RAM re-install. Arrggghhhh. syslog_Sept15

I'm on a slow boat right now switch off various bits of the system. I removed all bifurcation and it's still crashing. I removed Cache Dirs plugin which was pegging CPU, and its still crashing. Next, I'm switching off VM and Docker support. I'll grab a log after that if iot';s still crashing.

No Joy Still crashing frequently today. I've disassembled and rebuilt with some config changes and compromises and will see if that works. (I eliminate my bifurcated set up)

I ran this on all my array drives and it did it's clean up. I ran it a second time with no further issues addresses. Let's see if that made a difference. Thanks for taking a look. I appreciate it greatly.

Shares came back after a reboot, but significant issues persist. I've made an issue thread. 6.12.4 is crashing much more frequently, than 6.12.3 was (hourly rather than daily).

Hi folks. Ever since upgrading to 6.12.x, I've been having nothing but issues on my server. I recently updated to 6.12.4 as 6.12.3 was hard crashing every couple of days. Now it's crashing every couple of hours. Any thoughts, help or insights appreciated. There have been no changes to my system recently apart from an SSD swap to remove a flakey old drive, though the crash issues were occuring before this, and I did this change as part of troubleshooting. I've attached here my diagnostics pack as well as most recent ssysog. The behaviour I see is a standard start up, but after a few minutes or, at most, a couple of hours, I will observe a cascade of Kernel faults in logs that ultimately result in a system lock up necessitating a forced restart. The initial fault is never the same. I have run 2 passes of memcheck all day yesterday with no faults. I have several times checked SATA and power cables to drives and have verified all expansion devices are correctly seated. A fresh set of eyes would be appreciated as I'm at my wits end here. syslog unraid-diagnostics-20230912-1557.zip

updated to 6.12.4 as 6.12.3 was hard crashing every few days. (6.12.2 before it lost me days tracking down some kind of plugin corruption). Now all my shares have disappeared, meaning more hours of debugging and faffing around. What's happened unRAID? I've used it for years and it's been rock solid. Now nothing but issues and problems and hassle????

Reporting this encountered for the first time today on 6.12.3 Manifests as empty pages on the Array, Pool and Boot device tabs. Log just filled with this; Aug 8 13:45:08 UNRAID nginx: 2023/08/08 13:45:08 [error] 9990#9990: nchan: Out of shared memory while allocating message of size 19386. Increase nchan_max_reserved_memory. Aug 8 13:45:08 UNRAID nginx: 2023/08/08 13:45:08 [error] 9990#9990: *2742737 nchan: error publishing message (HTTP status code 500), client: unix:, server: , request: "POST /pub/disks?buffer_length=1 HTTP/1.1", host: "localhost" Aug 8 13:45:08 UNRAID nginx: 2023/08/08 13:45:08 [error] 9990#9990: MEMSTORE:01: can't create shared message for channel /disks Aug 8 13:45:08 UNRAID nginx: 2023/08/08 13:45:08 [crit] 9990#9990: ngx_slab_alloc() failed: no memory Aug 8 13:45:08 UNRAID nginx: 2023/08/08 13:45:08 [error] 9990#9990: shpool alloc failed

So I wiped my USB and reinstalled UNRAID. Copied my config directory and the issue persisted. (could not get GUI in safe mode). I then embarked on copying my config sub-folders and files one-at-a-time to see which one was the offender. A painstaking process of reboot after reboot. And would you believe - it was the VERY LAST FILE - the VFIO config. ARRRGGGGHHHHHH!! With a fresh UNRAID install + all my config folder except VFIO, which I rebuilt, I was finally able to create the ZFS array. It's chugging away now doing some test transfers so I can evaluate performance, but it seems somewhat on a par with my main array. I'll use it for a secondary backup for now so I don't commit any files I don't have copies of elsewhere. Thanks for looking.

In trying to troubleshoot my Pool configuration problem, I've been trying to boot my 6.12.3 server into safe mode bujt that's not working, so I'm hoping for a bit of guidance. Here's the link to the thread that gives rise to this one; Anyway, when I boot to SAFE GUI mode, server boots up OK, I log in on local display. but no GUI is displayed. Firefox says unable to connect. When I try to connect remotely, I get a connection refused in my remote browser, and SSH reports that key fo [IP] had changed and that host key verification failed. I've read up a bunch of the 'No GUI' threads, but found nothing helpful. Furtherest I got was finding NGNIX was not running. I started it and got a bad gateway. I restarted both it and rc.php-fpm and was directed to a blank page at /Tools/registration. (no other UI URLS would load either. I completely emptied my plugins directory, but had the same issues on reboot to SAFE mode. SSL is OFF So am stuck. Diagnostics attached. Current advice on other thread is to rebuild USB. Would like to no whats's wrong though. unraid-diagnostics-20230801-1602.zip

Im trying with both the built in GUI (monitor attached), as well as a remote connected Chrome browser. Same result in either case. i tried booting safe mode but there are problems there. In GUI safe mode, the browser won’t load the UI. In GUI or standard mode, a remote browser won’t connect, and even ssh fails to connect with password problems. I guess ive got to fix that problem before I can progress with my pools issue.

I can erase, but even after that the apply button remains disabled.

I am unable to configure new pools or modify existing ones under 6.12.3. What are the reasons for then 'Apply' button in pool configuration to be disabled? Discussion here;

No. If I create a new pool and add one or more devices, (any device) I cannot apply settings. If I access an existing working pool, make a small change, still the apply button is disabled. Essentially, I cannot make or edit pools, unless I'm willing to accept default settings for a new pool

Folks Really stuck here. In playing around with this, I find that I can neither set up & configure a new pool nor can I make any changes to existing pools while array is stopped as the APPLY button remains greyed out on the pool settings page. Diagnostics attached. unraid-diagnostics-20230731-1155.zip unraid-syslog-20230731-1054.zip

No joy. Didn't have an impact on the issue. I also remade the pool without the underscore character in the name, and reduced the number of drives to 3. But still the Apply button on pool editing remains resolutely disabled.

I'm running 6.12.3. I have a bunch of 'OLD' (ex-array) 4TB drives I've reinstalled to my server that I wanted to use to have a play with a ZFS pool. Here they are before I start. Unformatted, no FS etc. I created a pool (I have 3x other pools previously set up) and added the disks; So far, so good. Now, I click the link 'Zfs_pool_a' beside the first disk. (note, I DID NOT capitalise that name, I was careful to ensure I named the pool all lower case) I set up my file system type as ZFS and configured my settings; However, I cannot apply these settings. Note that the 'Apply' button remains stubbornly greyed out. In fact, it remains so regardless of what file system type I select. If I click 'Done' and start the array, it goes ahead and formats the pool BTRFS. I must be missing something obvious here, but what? A procedure? A setting? A checkbox somewhere? Pointers appreciated.

It's this one: Allegro Pro USB 3.0 PCIe 4-Port Computer Card - Sonnet (sonnettech.com) This is the elusive 4x Type A USB 3.0 port card with discrete controllers per port, meaning you can assign each port to a different VM, assuming your MB IOMMU breaks them out correctly.. I discovered that unfortunately it just didn't want to work reliably with my bifurcated PCIe ports. (Taichi X399 wouldn't always post). It works great when used in a standard PCIe port, but unfortunately did not meet my very specific needs. Hence sale. Bare card only. No packaging, though will be well wrapped for postage. Looking for $150 / €135 or close offer + postage at cost. Payment by Paypal (friends) or Revolut. I'm located in EU, (Ireland). I can post worldwide, but be aware you may incur import fees and charges. YMMV. I'll be listing on fleabay in a few days but wanted to give unRAIDers first shout as these are hard to find and if you need it, you need it.

Curious as to how you figured that out. Was it documented somewhere? Seems very counter intuitive. Anyway, I had the same issue when I upgraded from 6.11.5 to 6.12.2 and was super-frustrated until I found you post. Many thanks.

Thanks for help so far. Making good progress (and some of the above was down to non Unix line breaks in one of my files - D'Oh! One small item I cannot find an answer to..... Is it possible to direct link to a tab within a page? For example, if I click on a disk under 'MAIN', the link brings me to the first '<diskname> Settings' tab of the Device.page (/Main/Settings/Device?name=parity) Is there a URL that can be used to load up one of the other tabs in context of the parent page? For example will display the 'Identity' tab by itself (DeviceIdentify.page), but I'd like to use something like to show all the tabs, with that particular one selected.

Cheers That's exactly the folder structure I had, with no joy. However, when I simplified the file contents as per your suggestion, the tab appeared, so there must have been some error in the page itself. Now I can press on with this part. Appreciate it.

Looking here; I see these pages; Which are what (I believe) generate this; I want to add a 'driveTag' tab there. I've added this page inside my plugin; Which starts like this; But nothing shows. Trying to figure out what I'm doing wrong.

Thanks for the response. I guess I need 2 and 4 from your list. (I know how to do 1, don't want to do 3 for reasons you outline and not 5) I'm hoping to do 2 things; A new section (tab) within this second level page; And an additional icon, column or rollover (TBD) per line item for disks on the dashboard, and under each tab on the Main page where disks are listed;

Hi Folks, Trying to write a plugin and very frustrated with the scant documentation. I've figured out creating the skeleton, the plugin settings page, updates, saving out data and the like. Now, I want to add a page to the webGUI, and add functionality into some other existing pages, and cannot figure out how to best go about implementing such 'overrides' The idea is to add a new tab to /Main/Settings/Device?name=diskN to allow users add a note associated with the disk. Then, on pages where disks are listed (Dashboard, Main etc.), add an icon or rollover state to each disk to display the note. What's the best way to add functionality into /webgui/plugins/dynamix/ ? Can I dynamically create / replace pages in there from my own plug in? Surely not sustainable in the case of updates etc? I've had a good read through what's available in this forum, tried several search terms, and poked about a lot in existing plugins. The closest I could find to something I'm trying is @Squid's Docker Folder, which adds a button to a specific page in the UI. Are there any other plugins that add new pages or extra functionality to existing pages that I could look at? Any pointers appreciated. I'm just looking for a push in the general direction. Bonus question: how to gain access to the webGui design sub forum? Some of my searching led to threads there, but I'm not permitted to read them.

Not quite thin client, but I use these Av Access HDBaseT extenders to pipe VMs from my unRaid server to wherever I need them around the house https://amzn.to/3ZIEYZt The transmit up to 4K HDMI, USB and audio. On the client side I have a display, keyboard, mouse and speakers. They are a wonderful solution, very stable and flexible (PoE) on one end so only need to power one of them. The only real drawback I've come across is that they don't like high bandwidth USB devices. I can get my mouse and keyboard to work, but a set of USB Harmon Kadron Soundsticks are just too much for the setup. I'm have a USB soundcard plugged in on the server side and use the audio jack on the devices and that works great. I wrote up a blog post a few years ago on this. Given I'm still using the devices years later (typing on a VM now in my house from an unRAID server in another building), I'd say they are keepers.