salora

it is a standard AHCI SATA controler, but the connector is a mini sas port so you need the right cable to attach the sata drives like that for example https://www.amazon.fr/CableDeconn-SFF-8087-SFF-8482-Connecteurs-alimentation/dp/B010CMW6S4/ref=sr_1_1?__mk_fr_FR=ÅMÅŽÕÑ&keywords=CABLEDECONN+Mini+SAS+36+SFF-8087+to+(4)+SFF-8482+Connectors+with+SATA+Power+1m&qid=1585484930&sr=8-1

No sir I beg to disagree It's an embedded sas port that can handle 6 sata drives So of course all drives are connected to the main board and seen by unraid and not the raid controller

I know i've been using unraid for 4 years but still from 400mb/s to less thant 40 ! that is huge drop down

Indeed but like said already I tried with the embedded sata port of the motherboard with no better luck

So mooved everything to windows server and hyper-v with a virtual machine running the file server and smb shares the drives are in raid 5 on the perc h710 and I've more thant 400mb/s transfer rate so maybe a dell server is not made to run unraid I don't know

Yes I know it's not ideal, i'm trying to get hand on a h310 to flash on it mode I said i've done a lot of trials , there is a sata port directly on the motherboard I tried it an even that way with drives directly shown to unraid it's bad with or without turbo mode, reconstruct mode.. other than the perc there is dual xeon with 16 cores (and HT) , 128 Gb ECC ram so it's not a low end rig

config maybe, hardware it can't be otherwise i'll had the same issues with other OS

well okay I get your point I have these speed too while it's writing to RAM ( the dirty cache setting) but when it's filled the writing speed drop (like many others from what i've read) my "slowest disk" can perform at at least 120Mb/s on a raid array with windows or xpenology and not unraid ! that I don't understand I tried the DiskSpeed app to analyse and the average speed on each disk is much higher than that so I can't get why disks able to more than 150mb/s get down to 40 when putted in an array. I get Unraid is not made for top perfomance but still ...

reconstruction write mode is activate and no better performances i tried health monitoring is done through dell IDRAC so not an issue at all

Well I can say I made a lots of trials and I tried both solutions since, and what makes me perplex is the low writing rates on unraid and there are a lot of topics on that! drives are all sata 3 so with a normal array set up (drives and parity) writing speed don't go above 20mb/s to 40mb/s ish with turbo cacha and other solutions found (Disk Cache Settings..) on other topics reading writes are ok with the raid 5 on the perc and one big drive presented to unraid, no parity its better, writing speed up to 70mb/s. BUT with a the raid 5 presented to a windows Server OS well, writing speed is up to 120Mb/s , saturating a gigabit ethernet link, with the same drives and perc configuration with the sames drives on an Xpenology setup with a poor asrock j1900 motherboard writing speed is also up to 120Mb/s , saturating a gigabit ethernet link So I don't understand Unraid is basically a NAS OS for a start , why is the writing speed so low no matter what??

Greetings So I now have a dell t420 with the perc raid controler and 8 drives and i entend to replace my ryzen 7 unraid server with it With the perc you at least must do raid 0 virtual drives that will be presented to unraid I can't decide between 2 options I/ Make 4 raid 0 virtual drives with the perc and then have an array of 3 drives and one parity, unraid will then deal with data loss prevention II/ Make some raid 5 or 10 with the perc and an array with no parity and have the raid controler deal with data loss prevention. I can't decide and don't know what will be best so if anyone have advice?? thank you best regards

Greetings everybody, so I have the same issue as others here, so I stayed with 6.7.2 hoping that maybe 6.8.1 would solve this unfortunatly not, so besides the workaround (that may induce a loss in performance), will this be fixed ? or we shall wait for unraid 6.9? thank you regards

Hi everybody, back to seek knowledge for fine tuning of my unraid server I ran a cpumark test on a windows VM with all cores passed it barely hit 9000, ryzen 7 2700 should it around 15000 cpu governor is on performance and turbo boost on so I tested this command cat /proc/cpuinfo | grep "MHz" to see the speed of the cores it never goes higher than 3.4Ghz even with intense stress more weird is if I run this command directly on the unraid gui selecting option 2 at boot (and not from a remote navigator) it reaches 4.1Ghz So please tell me if the score on cpumark is normal and if I can do something to help the cpu to reach the turbo frequency thank you best regards

Hi guys, so i've just built my New ryzen unraid server (see signature for details) I moved from an I3 unraid server + an i5 6600k gaming machine to this 😃 Everything is working great, i've succesfully passed my 2 graphic cards to vm and the gaming vm has quite good performance 😁 Now it's time for optimisation so is there mandatory BIOS settings to apply? And does a Windows vm use the Turbo capability or the cpu or is it better to overclock? Thanks best regards

thanks no there was no passthrough on the one that might be on autostart but I figure this out by deleting the libvirt image and rebuilding all my vm's so problem solved I would say last time I had a server upgrade it was from an I3 to an I5 and it went smoothly this time it's from I5 to Ryzen so it's a bit more complicated

So I'm able to start libvirt and create VM's if I do not load my old VM's by changing the default storage path So my guess is an "old" VM is causing the trouble maybe I did let one in autostart So new question how do I disable an autostart VM without starting libvirt? thanks

Same in safe mode 😥

Feb 23 20:18:28 Serveur kernel: md: disk0 read error, sector=972118240 Feb 23 20:18:28 Serveur kernel: md: disk1 read error, sector=972118248 Feb 23 20:18:28 Serveur kernel: md: disk2 read error, sector=972118248 Feb 23 20:18:28 Serveur kernel: md: disk3 read error, sector=972118248 Feb 23 20:18:28 Serveur kernel: md: disk4 read error, sector=972118248 I got this repeat after trying to start libvirt And I attached another diag file from after the crash serveur-diagnostics-20190223-2019.zip

So I'm struggling since hours I had my server running for month on intel system with no issues I've updated to a ryzen system with ryzen 7 2700 , gigabyte ga ax-370 gaming 5 system online, array is working, docker aswell all drives seems good with no smart errors parity is valid BUT when I try to start libvirt it cause the array to crash with drives going faulty and VM and isos storage being unavailable. all drives then go to unassign devices I really don't understand what is happening I've updated from 6.6.6 to 6.6.7 with no success So if anybody can help it will be much appreciated serveur-diagnostics-20190223-1956.zip

ok so passthrough is working now too Docker seems very slow now , weird and I still don't know why I have this write error broken pipe ! but it's working now so big thank you for helping me best regards

the only file I did change was the "go" file for plex hardware encoding so I start with that I putted this "#Setup drivers for hardware transcoding in Plex modprobe i915 chown -R nobody:users /dev/dri chmod -R 777 /dev/dri" Back to the original one its working with gui even if I still have that weird broken pipe error so now I'll try without the gui edit : it's working too, still the broken pipe, now it's time change the syslinux conf to enable the passthrough

so it is working like that What do I do next, do you need to see some files?

I will try ok moreover in safe mode it's not working so i would say that the plugins are out of this

anyway, i paid the license so... I hope that with the logs someone will help me figure this out

If you never stop it , it can last forever you can't upgrade but it's the same result if i cannot update it because of a bug Moreover it's not some old rig, but kabylake intel cpu with an updated motherboard with the last microcode