Setup:
Chassis: CSE-216 with SAS216A backplane (non-expander)
Motherboard: Supermicro X9DR7-LN4F with very latest BIOS
RAM: 64gb Samsung ECC Registered on the HCL
CPU: E5-2620V2
Storage Controllers: onboard LSI-2308 - works perfect
2x Intel RS3WC080 - LSI 9341-8i
Drives: 9x Samsung PM853T
So here's the scoop:
I'm using a group of 6 of the Samsung drives in a StorageSpaces RAID0 for testing. No critical data on them. I'm running the controllers in JBOD mode and they have the latest Intel firmware on them. Here's the rub: write performance is HORRIBLE. I loaded up MegaRaid Storage Manager to look for errors and the system was spamming dozens of "Unexpected Sense: Power on, reset, or bus device reset occurred" errors. I checked the controller's temp and they're 55C give or take 5 degrees. I checked the cabling and I'm using brand new LSI-branded MiniSAS 8643-8087 cables. There is no strain on the cables and they're properly attached to both the cards and the backplane.
I tried hard-setting the link speed to 6G and that removed most of the unexpected sense errors. I started backing up the data to an off-system backup drive I have and got a few Unexpected Sense Information CRC error detected messages. I haven't tried any more writing because I don't want any dataloss but this is getting kinda nuts. The backplane including those exact slots works perfectly with the onboard LSI-2308 controller so I'm pretty sure it's not defective. Any hints here? Should I try flashing the RAID cards to an LSI factory 9341-8i ROM? Will doing so remove the RAID5 and Cachecade keys that these cards came with? Any way to back those up so I could recover them if so?
Sorry, I'm at a bit of a loss here!
Chassis: CSE-216 with SAS216A backplane (non-expander)
Motherboard: Supermicro X9DR7-LN4F with very latest BIOS
RAM: 64gb Samsung ECC Registered on the HCL
CPU: E5-2620V2
Storage Controllers: onboard LSI-2308 - works perfect
2x Intel RS3WC080 - LSI 9341-8i
Drives: 9x Samsung PM853T
So here's the scoop:
I'm using a group of 6 of the Samsung drives in a StorageSpaces RAID0 for testing. No critical data on them. I'm running the controllers in JBOD mode and they have the latest Intel firmware on them. Here's the rub: write performance is HORRIBLE. I loaded up MegaRaid Storage Manager to look for errors and the system was spamming dozens of "Unexpected Sense: Power on, reset, or bus device reset occurred" errors. I checked the controller's temp and they're 55C give or take 5 degrees. I checked the cabling and I'm using brand new LSI-branded MiniSAS 8643-8087 cables. There is no strain on the cables and they're properly attached to both the cards and the backplane.
I tried hard-setting the link speed to 6G and that removed most of the unexpected sense errors. I started backing up the data to an off-system backup drive I have and got a few Unexpected Sense Information CRC error detected messages. I haven't tried any more writing because I don't want any dataloss but this is getting kinda nuts. The backplane including those exact slots works perfectly with the onboard LSI-2308 controller so I'm pretty sure it's not defective. Any hints here? Should I try flashing the RAID cards to an LSI factory 9341-8i ROM? Will doing so remove the RAID5 and Cachecade keys that these cards came with? Any way to back those up so I could recover them if so?
Sorry, I'm at a bit of a loss here!