ChatNoir

Ok, no problem, I just wanted to be sure I was alright before starting configuring several dockers and stuff. Thank you Squid.

I'll try to clarify my reasonning. On 6.8.3 and before you could use cache to: cache data before transfering it to the array: Yes store data that you used often and wanted a fast access to: Prefer or Only for all elements where speed was not important could be set to: No With 6.9 and the multiple pools, I understood that those different tasks could be splitted on different pools that could be named cache or whatever. The end goal for me is to have two pools : cache : for data transfer cache local_data : for dockers, appdata and the likes For the moment, I am waiting for more stuff to advance the build. The array is not in its proper config and I did not load all the data I want to on it. So I am just playing around, trying to set things up, making tests before finalizing everything. I do have a SSD, but only one available for the moment, so I focused on the dockers to have the test setup running, not the transfer cache. I am not at home to make screenshots of the details of the shares, but as shown on my first screenshot, I selected Only and the only pool available local_data, not cache for the shares in question. All media etc shares are set to No. It is possible that I misunderstood how to setup my pools and/or shares as I only have unRAID running for a few days but it seems to me that the plugin is only looking for a specific pool named cache and does not consider that I do indeed have a pool, but with a different name.

Hey, I am just setting things up while I wait for the final components. I just got 3 errors and the plugin is not happy with me. I looked in this topic for the last few months of posts without succes. With the opportinity now in 6.9, I wanted to split cache and dockers/VMs/appdata so I first create a pool named local_data and I was planning to add another pool named cache later on. I chose not to use any caching for the moment on media and personal shares and set every other one to local_data Still I got this errors and I am not sure if I did something wrong or if the plugin is only searching for a pool named cache. Do I have do delete the *.cfg files as suggested from the key ?

That's what I tought initially, then I re read : Did you tried the disk directy on a SATA port ? This board or another computer ?

Did you try disableing both v1 and v2 as suggested by @Xaero ? I would guess min protocol = SMB3 both I am no expert. ^^

As I don't have Kapton on hand, I won't have an array for a few days and I will wait for the array and the initial data transfer before assigning the cache. So, not much to do but better be productive in the mean time, what can I do to prepare the server ? I was thinking about : - configure notifications - install and configure some plugins (unassigned devices, Preclear, Fix common problems, etc.) - make sure the reported temp are OK Any advices ?

Wouldn't Home Assistant be the all in on you are looking for ? Not something I use but I do hear about it very often. Seems to match your requirements though and much more. 😁

I cannot answer for LT but I doubt any of this (ZFS, multiple arrays, etc) will be integrated in 6.9. 6.9 is in public beta and I think that they are fixing the last bugs before RC. Those elements will be probably part of a future release, probably even several different release given the amount of development for a small team.

Alright, it's alive ! Some BIOS issues during the weekend prevented me to progress much, but it is now OK with the help of a Supermicro rep. Booted up during my lunch break and the HDD I plugged is not recognized, only the SSD. It was working fine in my USB dock so I suppose that I have to put Kapton on the 3.3V pin of the drives ?

You might want to try unRAID 6.9 beta, a lot of drivers for recent NIC have been added.

What is your new hardware and what version of unRAID are you running ?

So, with all the previous discussions I did more researches and I went with a more recent generation of board and CPU. I also shifted from server board to workstation since I the specs and prices were more in line with what I wanted. I got : MB : Supermicro X10SRA-F (IPMI and VGA output, plus it was cheaper than the regular SRA boards ^^) CPU : Xeon E5-2697v3, I doubt I'll ever lack horsepower but I could upgrade for a v4 in a few years if needed RAM : 4x 16GB Samsung ECC RDIMM GPU : I will make some tests to see if and how I can use the VGA from the MB and I will try to use my old GTX770 for encoding Drives : 2x shucked WD140EMFZ (will be 1 parity and 1 data) and I'll empty my Syno once every thing is running smoothly SSDs : one WD Red 500GB for the moment, I will probably double it to have some RAID in a few weeks Thanks for your help, hopefully everything goes well and I will not have to ask any questions on the forums (yeah ... not 100% sure here ^^)

Oui, tu peux faire du raid sur ton cache comme maintenant. Mais le multipool ouvre beaucoup de perspectives. Le cache de transfert classique / un pool pour les Dockers / un pour les VM, voir par VM, etc. La complexité et le manque de flexibilité de ZFS est un des gros éléments qui m'ont éloignés de FreeNAS (et la différence par rapport à la communauté unRAID aussi ^^). Mais je j'ai un peu d'espoir quant à l'avenir de ZFS. J'ai lu qu'ils travaillaient sur la réduction, suppression de VDEV. Je pense qu'on est encore loin que ça sorte ou que ça ait la souplesse de unRAID mais d'ici quelques années ...

Pas mal de points ont été abordés, je vais essayer de donner mon avis de novice ici et là. #3 : Pour la quantité d'espace disponible tu peux regarder du côté de SHR1/2 en effet, même si la vitesse de lecture sera apparement moindre. Par contre, il me semblait qu'il n'y avait pas forcement de reconstruction de parité à l'ajout d'un nouveau disque si on avait fait un preclear sur le nouveau disque. Il rentre dans l'array avec tous ses bits à 0 donc sans remettre en cause le calcul de parité. #5 : concernant le cache, il peut avoir plusieurs fonctions : servir de tampon en écriture sur l'array héberger des données/images pour les dockers/VMs afin d'avoir des accès rapides d'autres usages auxquels je ne pense pas forcement (<- noob) A noter que la 6.9 actuellement en Beta permet de gérer plusieurs pools de disques en plus de l'array, pas uniquement le cache. Les pools de disques (cache notamment) peuvent être utilisés avec un disque ou avec plusieurs en RAID pour plus de sécurité pour les infos des dockers par exemple. #7 De ce que j'ai compris, c'est réalisable en USB, ou branché en SATA avec Unassigned Devices non ? #8 d'accord avec Baloo, tous les disques comptent comme indiqués sur https://unraid.net/pricing : Attached storage devices refers to the total number of storage devices you are allowed to have attached to the server before Starting the array, not counting the USB Flash boot device. There are no other limitations in the software based on registration key type. Pour ce qui est des vidéos, je te conseille SpaceInvaderOne sur Youtube. Il y a énormement des tutoriels intéressants.

Bienvenu Sylvain. J'imagine qu'il y a des limitations sur les comptes récemment créés ou avec peu de messages pour éviter le spam.

I see that you have a x16 PCIe slot, is it occupied ? If not, you could use a PCIe to eSATA or SAS in the future.

It would help the guys give you an accurate answer if you could provide your diagnostics in your next post.

You did not use a parity drive ? Or you are still doing the initial file transfer ? Also I see that your drives are very hot, so I would suggest to adapt your cooling solution to do something about this. Finally, half of your drives show SMART errors, you should look into it. Though it might be because of the temperature.

Si j'ai bien compris, on peut "valider" la valeur actuelle pour remettre le disque en vert. Comme ça si ça réapparaît on peut plus facilement déceler le changement d'état.

C'est ce que tout le monde fait en général, mais j'ai remonté mon précédent PC jeu en bureautique pour ma mère et j'avais donné tous les autres vieux machins à une association. C'est dommage de ne pas se faire les dents à coût 0€, mais il faut voir le positif, je commencerai avec du matos plus performant. Assez parlé de mon matos sur ton sujet. :-)

ebay est mon ami et surtout je voulais des ports SATA sur la CM + des lignes PCIe utilisables. Du coup, j'ai un peu cherché et voilà.

Il me semble que les erreurs CRC sont plutôt des erreurs dans la communication, soucis de connexion ou de contrôleur. Tant que ça n'évolue pas ça devrait aller. Mais j'attends toujours ma mémoire pour installer unRAID pour la première fois alors tu vois bien mon niveau d'expertise. 😁

No weird setting in the BIOS that handles both CPUs differently ? Did you try you HW on another OS ? Say an ubuntu live and check what is the base speed in that situation ?

Après, sans aller sur un monstre, beaucoup de gens utilisent le Node 304 de Fractal Design. Tu devrais pouvoir trouver des tonnes de vidéos sur Youtube.

It really depend on whether or not you are using the optical drive right now. If the answer is never ... why bother ? The answer for my usecase would be : - the 128GB on 8th port for occasional use - cheap usb drive in the array (if you never use it anyway ...)