LammeN3rd

did not think about trim.... I agree with you on this one!

might be stable but would you choose them over a cheap SAS2008?

Yes! the Z77 chipset is connected by a DMI bus to you cpu (4x PCIe) and the ASMedia is connected to the Z77 running an SAS 2008 on an 4x PCIe (gen2 because the SAS2008 does not support GEN3) gives you max 250MB/s per disk with 8 disks in parallel. if you only connect HDD you will only get there with very big and fast drives! Besides speed SAS controllers (at least from LSI) are way more stable reason enough to stick to those since they cost very little nowadays

If you can, don't use you motherboard SATA, these are horrible! (no matter what board you have) sata drives are fine for storage SATA controllers however are not made for parallel acces, in the best case they cripple your performance in the worst case you have drives dropping out putting your array in danger!

the Link speed of these controllers is per disk (when direct connected to the controller without expander) so if you connect both 6Gbit and 12Gbit directly to an H330 both will run their full speeds. with expanders it's an different story and without DataBold it will run on the lowest speed. DataBold is as far as I know its an LSI tech so you need an compatible LSI expander and LSI controller From the graph you should be running in Config 3 so you get max BW to all cards! if I remember correctly I'm running the SAS35X28R SAS EXPANDER connected with 8 lanes to the H330 (both ports) this is my full config: Dell Poweredge T630 | unRAID 6.6.6 - Pro | 2x E5-2640 V4 | 128GB ECC REG | PERC H330 12Gbit SAS (HBA Mode) | 18x Hotplug disks with 12Gbit SAS expander Backplane Array 72TB with Dual Parity | HGST Ultrastar He10 10TB [ P1, P2 ,D1, D2, D3 ] | Seagate IronWolf 6TB [ D4, D5, D6, D7, D8, D9, D10 ] Cache 1920 GB RAID1 | 4x Samsung PM963 960GB NVMe

If you enable Turbowrite you need full bandwidth of all disks, read/write to a share is max 1 disk of throughput for parity checks or rebuilds you need full bandwidth as well. the test I did was a parity check on the disks combined with a BTRFS scrub on the 6x SSD cache (since moved to NVMe)

An SAS expander is like a switch, it's usually found on backplanes connecting to a gig amount of drives. a 36 port expander can connect to a SAS HBA (SAS Card) with 4 or 8 lanes (ports) and the rest of the ports connected to disks. the disks will share the bandwidth, with older SAS tech this could give performance bottlenecks but with SAS 12Gbit and Databold the only way you can run into that is with very fast SSD's My server has an 26 port expander 8 for the connection to my Perc H330 and 18 for the disks. since all my disks are 6Gbit SATA even 18 drives will not create an bottleneck on the 8x 12Gbit lanes from the H330

If you combine this card (or another SAS3008 based card) with an compatible LSI 12Gbit SAS expander you can add way more drives without losing performance. older cards and expanders would run on the lowest link speed but these support DataBolt technology so you can run multiple 6Gbit drives on an expander without losing any performance! I did an test a while ago with 18 drives connected (12x HDD and 6x SSD all SATA) and got ~5.5GB/s the full bandwidth of 8x 12Gbit SAS is not possible since the 8x PCIe bus is not fast enough

The H330 usages the megaraid_sas driver and works oud of the box! I use it for about a year now in Unraid (HBA mode) never had any problems with smart (or anything for that matter) It’s wicked fast!! Sent from my iPhone using Tapatalk

The Perc H330 works great, never tried to flash it to the IT firmware. the Dell HBA330 is even better (same HW but with IT firmware out of the box)

Dell Poweredge T630 | unRAID 6.6.6 - Pro 2x E5-2640 V4 | 128GB DDR4 ECC REG PERC H330 12Gbit SAS (HBA Mode) | 18x Hotplug disks with 12Gbit SAS expander Backplane Cache 1920 GB RAID1 | 4x Samsung PM963 960GB NVMe CPU's are on power save profile, this does lower performance but 2x 10 Core Xeon has more than I need anyway and running on lower frequencies increases the perf/watt massively. Allocation method: Most Free Set all disks to 15 min spindown except the last drive in the system (if more than 1 disk is empty exclude that for share) that has no spindown and gets all new data. this keeps the disk with all recent stuff online so only slow spinup's for older stuff.

Really love this option!!! update went great!

I've been running UNRAID on a T630 for about a year without any issue's, created an USB stick with the unraid tool. both UEFI and Legacy boot work without issue and the PERC H330 in HBA mode is great for performance! I suggest resetting your bios settings and maybe update your bios to version 2.8 Good luck!

The latest build (153) did not come online on my server, nothing in the logfile after Server startup complete. Host name is.... reverting to build 152 and it's back online again. No idea what's wrong if there is anything specific I can test / logfiles I can get please let me know!

besides the new font sizes I would love to be able to change / invert the background and text just below the banner in the dark theme! I've already made a black banner

Check your profile page under activity to see more than the first 3 lines refresh the page while filming it (it shows the rest for a split second) the new new logo I was taking about, I just found the url for the new forum

ohh wow! really like the new logo! won't spoil the party by posting it already but it's not hard to find

There is more.... To bad its not about the next version of unraid.

Yes it does, only PCIe link speed though (with 8x 12Gbit SAS thats the bottleneck anyway)

yeah I need more drives, my controller is nowhere near maxed out! (Dell H330 in HBA mode with a 18 disk expander backplane)

Beta 5 recognizes all of my drives

the Parity check on my Array with 10TB disks (7.2K) takes about 16 hr 52 min without the 6TB drives that would be just over 14 hours These big drives are quick, as long as your controller can handle it

I would love to see additional information in the Parity-Check History file, adding the unRAID version and the number of disks in the array. Besides the extra columns it would be great to have other significant array changes to the array in there as well time-stamped and all persistent after reboots. I'm thinking of: Unraid Software updates (from - to version and with date and time) Disk changes: disk number X with Identification Y replaced by Identification Z Disk Additions: disk number X with Identification Y added to Array Unclean shutdown with Date and Time I'm sure others with actual experience in trouble shooting unRAID specific issues than me can chime in to make sure we have the most valuable info without creating a needle in a haystack situation! Having this info will go a long way in trouble shooting all kinds of issues where order of events is key! Relying on human memory to figure out if and when certain events have taken place is not accurate or lossless option for trouble shooting issues

You need Bios support for PCIe bifurcation without only the first NVMe SSD will work. Bios support for NVMe is an different thing and would be needed if you want to boot from your NVMe drive, since unRAID boots from USB NVMe boot support is not needed!

I had the same in my Dell Server (T630 with dual Xeon E5-2640 V4) and it turned out that I set it to manage its power/turbo by the BMC not the OS, changed the setting and now the unRAID does the work! I've no experience with HP servers but I'm sure others have