12_purple_giraffes

So 2 or 3 weeks ago i had to move over to a new hardware and ever sines iv been getting some read errors on two drives. I did a long smart test on both drive and no errors. The odd thing to me is that errors seem to happon overnight then the drives should be spun down as this is mostly a plex server and its only me that uses it. Can anyone help to understan what is going on her. I used to have all drive pluged in the motherboard but had to go with a HBA as not meme bored with 10 sata parts these days. I tempted to just go and get 2 x 5 ports: JMicron JMB585 and do away with the HBA. The hardware change was due to boot loop with Ram error codes. Old system msi x99a gaming pro carbon CPU E5-2630L V3 Ram 4 x 4gb New system Gigabyte B550 GAMING X V2 AMD Ryzen 5 5500 Corsair VENGEANCE LPX DDR4 RAM 16GB (2x8GB) 3200MHz CL16 IBM M1015 in IT Mode carbonblack-diagnostics-20230920-1419.zip

So i did a convert to raid 1 that seen to fix my cache pool but my docker have stop working getting "Docker Service failed to start." in the bocker tab they had stated to work before arfter a shutdown and restart. New Diagnostics attachned. carbonblack-diagnostics-20230822-1912.zip

Im gettting To my knowledge im im not using pool balance. I have just had to change Motherboard, CPU and Ram also whent from having all drive attached to Motherboard to to a mix of HBA and Motherboard. The pool SSD are attachhed to the Motherboard. On first boot one of the pool SSD was missing (Not pluged in) i did a clean shutdown pluged in the drive and booted up. Also the pool is 1tb it should be 500gb using 2 500gb SSDs The hardware change was due to boot loop with Ram error codes. Old system msi x99a gaming pro carbon CPU E5-2630L V3 Ram 4 x 4gb New system Gigabyte B550 GAMING X V2 AMD Ryzen 5 5500 Corsair VENGEANCE LPX DDR4 RAM 16GB (2x8GB) 3200MHz CL16 HBA INSPUR 9207-8i All the drive on the HBA have been pickup. Also note that my Dockers are not running getting carbonblack-diagnostics-20230821-1906.zip

Yep using my "Network Share" that has a cache If i understand correctly setting Tunable (md_write_method) to "reconstruct write" would be Turbo write but all the drives need to be spinning i spin down my drives after 1 hr as its mostly used as a plex server not shore that would work great for me. My drives have only 400 to 800GB per drive free using 8TB drives.

So i recently upgraded to 2.5 gigabit networking and have noticed my transfer speed when writing is only hitting around 130MB/s read is 210-230MB/s (To and from the SSD cache) When writing i have one core on the server at 90% or more then writing is this a problem. Server M/B: MSI X99A GAMING PRO CARBON CPU: Intel® Xeon® CPU E5-2630L v3 @ 1.80GHz Memory: 16 GiB DDR4 Storage: Cache: 2 x Crucial MX500 500GB Array: 6 x Seagate 8TB IronWolf NAS PC CPU: AMD 5800X M/B: Gigabyte Aorus x570 ultra Memory: 32 GiB DDR4 Storage: Samsung 970 Evo 500GB (OS) Samsung 950 Oro 256 GB (Doc) Sabrent Rocket Q 2tb (Steam)