dada051

Maybe Multiple Arrays with the possibility to have to types of Array, ZFS Array and Unraid Array ? I don't see the need of a ZFS Cache pool, but a ZFS Array, yes. So I think multiple arrays is a good solution, and the ability to create ZFS Array can came with that

It's not limited to 30 drives, only the array. But I'm ok with the fact that you expect unlimited drives in the array.

lsof seems to not going away anymore : https://forums.unraid.net/topic/114162-100-cpu-usage-lsof

Look for lsof in htop. cf https://forums.unraid.net/topic/114162-100-cpu-usage-lsof You have to wait for a few seconds.

Sorry, but you didn't find anything more than we already discovered. The regex is use to count the number of open files (streams) in the dashboard that are not Apple related. The thing is that unraid should stop the process when the websocket connection has no more subscribers (kind of, I'm not expert). It's not the case.

I use a cache pool with raid 1 btrfs. Performance are the one I expect for a 2.5Gbps ethernet and SATA SSD.

If you never tried or bought Companion, only Array, it's not the good thread. I think an Array thread should exists

I checked for more details, I seems not working when path is a sub path of an always used path. I.e: /mnt/user/appdata/swag is used by swag, but the plugin offers me to delete /mnt/user/appdata/swag/log/nginx/ that was used by goaccess (that I deleted because it's now integrated as swag dashboard) Edit : when I say "used" it's that the path is in the template and so subpath is not referenced in the template, but it's used

Is it possible to check if offered directories are always used by currently running templates? I have some path used by removed dockers that are always used by others dockers and so they don't need to be deleted

I don't understand why you talk about performance array. Cache pool should be better for that use case. You can use RAID 10 to even improve performance.

I just found that the WebUI makes requests to wss://xxxxxxxxx.unraid.net/sub/cpuload,update1,update2,update3,wireguard update1 return something like : the percentage numbers matches with the Memory part of the Dashboard, the RPM numbers matches with the Airflow tab. The 0s are the numbers of streams (last column) of the Shares tab of the Dashboard.

Same problem here. I found that the lsof command is started from this file: /usr/local/emhttp/plugins/dynamix/nchan/update_1 I think it's related to the WebUI, but I'm not sure.

Same problem here. Not sure to understand why there is a correlation between lsof and bad HDD. @Osiris did you found anything else?

I'm not sure array with parity performance is better...

You can do this with cache pool. I think if we need multiple arrays, it's because we need multiple arrays with parity. If we don't need parity, we can do unassigned devices or cache pool. For SSD, the same (as unraid parity is not great with SSD). Multiple arrays allow to reduce the mix of drives like 5x 2TB 5400rpm 2.5" laptop drives and 4x 14TB 7200rpm 3.5" NAS class drives for example, and it allow "turbo write" on theses drives without waking up all the drives

Create an cache pool with raid 1 (2 disks, protection by mirroring disks) , unraid parity is not good with SSD/trim

Hello, my assigned devices plugin loads indefinetly in the main dashboard. Just updated now. Any helpful advice? Edit: add log, probably just after the update Apr 11 14:34:05 Tower nginx: 2022/04/11 14:34:05 [error] 13565#13565: *9508045 FastCGI sent in stderr: "Unable to open primary script: /usr/local/emhttp/plugins/unassigned.devices/UnassignedDevices.php (No such file or directory)" while reading response header from upstream, client: 192.168.1.74, server: , request: "POST /plugins/unassigned.devices/UnassignedDevices.php HTTP/1.1", upstream: "fastcgi://unix:/var/run/php5-fpm.sock:", host: "3726deb347de5f61998eb4fe9f92141b2169d5ce.unraid.net:5021", referrer: "https://3726deb347de5f61998eb4fe9f92141b2169d5ce.unraid.net:5021/Main"

6.10.0-rc3m?

I personaly add an DNS entry in my pi hole to resolve my Vaultwarden domaine to the local IP of my server. An alternate solution is to edit the c:/windows/system32/drivers/etc/hosts

New base for unraid 7 ? 😇

If I want my dockers on the "main" unraid, how I can easily (and without performances loss) share my unraid shares that are on the hosted unraid. And multiples unraid are not so easy compared to one unraid with multiple arrays.

But I don't want to isolate my shares

I'm ready to pay one licence by array if needed

I added a small JS code in a user JS extension : window.onload = function(){ document.getElementById("filter").focus(); }; maybe it's something that you can integrate directly?