iamr00t
Members
-
Joined
-
Last visited
-
Periodically, I've found if I leave a web browser tab up looking at either Dashboard or Main or Docker tabs (6.x and now 7.0 since beta 1), there's a good chance I can get the whole GUI to crash. If you google "Unraid GUI unavailable" or "Unraid nginx crashes" or "how to restart unraid GUI", there are dozens of posts from people experiencing the issue, usually coming to the conclusion that they need to restart the server to get the GUI back. nginx is throwing the error, so many people have recommended running the following to restart the entire webserver: /etc/rc.d/rc.nginx restart However, this has never fixed my issue (and I haven't found any obvious crashes in my logs), so I've ended up rebooting, which is tedious and disruptive. However today, deep in another forum post, I found this process (run by ssh'ing in to the server and running as root), which corrects the root cause for my GUI issues without a reboot - a PHP crash: /etc/rc.d/rc.php-fpm restart Now I know to look for PHP errors the next time this occurs, and can hopefully help contribute to identifying root cause... but in the meantime, I'm hoping this post bridges the gap on web and forum search results to help others as well.
-
nginx is throwing the error, so many people have recommended running this to restart the entire webserver: /etc/rc.d/rc.nginx restart However! Many thanks to another forum post, I found this which corrects the root cause for my GUI issues - a PHP crash: ssh in and switch to root (or use sudo) to run: /etc/rc.d/rc.php-fpm restart
-
This just started for me as well... entire server becomes nearly completely unresponsive, I can't do anything in the UI except login. I can usually SSH in but most things don't return. Every time it happens, and I eventually get in and run a ps, I can see several old random hung transcode sessions from my Plex container. It doesn't seem to be filling RAM (I do use /dev/shm for transcodes) and I'm using a Tesla P4 card with the latest nVidia driver. I am going to try downgrading to "Production" (from version 565 to version 550) and see if that helps... but right now the only fix has been to power cycle the server because poweroff and reboot at the console don't ever do anything (I can trigger these remotely as it's an Epyc system with IPMI). I've been rebooting about every 3 days for 3 weeks now. Edited to add: when I do experience this, manually trying to kill the transcode PIDs does nothing; commands trying to stop the docker container or daemon just hang.
-
I've recreated this issue twice as I expand my primary array, previously in 6.12.8, now in 6.12.10: Purchase new disk(s), slot into backplane Drives are recognized under Unassigned Drives Start Preclear on some/all drives, wait for drives to complete Preclear. Optional: Purchase and add more new disks to backplane Once Preclear is complete, stop array to add drives Add drives, start array, this succeeds as expected Tick the "Yes, Format" and click Format to officially add the precleared drives to the array Formatting begins, and because I am impatient, I go down and start preclear on any new Unassigned Devices I may have installed but not yet cleared Formatting will now silently stall until the Preclear(s) are complete (25+ hours for 18TB drives). I recognize this is likely intended or untested behavior, and an edge case. And now that I've found it, it's easy to avoid. I think my low-priority bugfix suggestion would be: have Unassigned Devices grey out Preclear while a formatting operation is occurring, as Formatting should only take a few minutes for Precleared drives to finish on an array, and/or have a warning note on Format button to please wait before starting additional Preclear functions. You can see what happens in my attached syslog: I start Formatting 3 precleared drives, click Preclear on a new drive (not logged in syslog, afaict), then the Format process just stalls completely for 12 hours. I then realize the Preclear is probably causing me grief, stop the preclear, and the stalled formatting completes quickly as expected and the drives are available in the array. (Array drives: Seagate 18TB Ironwolf Pro) hoard-diagnostics-20240411-1619.zip
-
Since this thread is still popping up at the top of Google searches for "unraid array size limit," I thought I would update the details based on my experiences using Unraid 6.12.4+ : The direct answer of "no, you can't have parity drives in pools" is technically correct if you are referring to "parity drives" in the way the array uses them. However, You can - and should - configure pools with redundancy as allowed, or striping as preferred. I have a pool just for cache without redundancy that is btrfs striped for optimum performance, and an SSD pool configured as RAIDZ.1 ZFS for redundancy on my appdata / docker config setup. The addition of ZFS in particular has opened up a ton of options going forward, though the share confguration and mover config not allowing easy pool-to-pool data migration is the biggest hiccup. Of course this can be overridden by creative script work, so YMMV.
-
@Eurotimmy Thanks again for creating the template; how did you get this configured on your local instance?
-
This is a great idea and I'm glad to help flesh out the community knowledge on this. I've installed the container using the new unraid template, but I'm having some difficulty in configuration. The template doesn't have an entry for a config directory, and no config.yml is created under the /mnt/user/appdata/romm/ directory tree. Modifying config.yml is the way ROMM usually works, for things like mapping custom library paths to supported platforms, however if I execute a shell in the running container (which means looking at the underlying disk from the docker daemon, not the user-modifiable appdata directory), I only see /backend/config/tests/fixtures/config.yml ... which is marked internally as a test/sample file only. Can someone fill me in on the missing piece here? Is the "test" config.yml actually live, so if I create a manual mount path for /backend/config/tests/fixtures/ in unraid's docker config for romm, it will allow me to create and modify my own config? Should I create a configmap for config.yml to mount at /backend/config/config.yml in the container, which will override any default configuration? Thanks in advance, I will be glad to help troubleshoot and share anything that can help here.
-
Have you tried 6.12.5rc1? They claim to have ZFS import working.
-
Just purchased a license and set up my server... perfect timing, as this feature makes my dashboard much more readable in one screen! Thanks!
