AOC-SAT2-MV8 and large drives

[bobmagnuson]

I've got an older system, one of the old TAM computers from years ago.  It has the AOC-SAT2-MV8 PCI-X controllers.

 

I'm running a mix of 2 & 3 TB drives without a problem, though at boot, the 3TB drives show at 2TB but unRaid knows better and recognizes them correctly..

 

I added a WD 8TB Red drive and unRaid saw it immediately.  I tried to preclear, but it crashed the system and with the drive installed unRaid will not boot.

 

I know this is ancient technology, but is anyone else running larger drives with this controller??

[BobPhoenix]

Never had 8TB drives but I did use 4TB and they were no problem.  But I don't think drive size matters if unRAID is able to see the whole drive when connected.  Are you using unRAID v6?  Because the SAT2-MV8s uses Marvel chipsets and those are know to cause problems for some people but not others on unRAID v6 (I.E. it can be hit and miss).  Also there are indications that specific firmwares/ATA levels can cause problems and others work fine.  So in that case it might just be the case that 8TB drives are not supported on your controller if using unRAID v6.  I have never seen anyone (until now maybe) report any problems with unRAID v5 and Marvel controllers but v5 is 32 bit so the drivers would be different.

[bobmagnuson]

Thanks for the reply.  I'm running 6.3.5 so am current.

 

Maybe it's time to upgrade the hardware. 

[bobmagnuson]

This is the disk log when adding the drive while unRaid is up, but the array is stopped...

 

Oct 10 18:33:22 Tower kernel: ata8: SATA max UDMA/133 mmio m1048576@0xd8200000 port 0xd8238000 irq 24
Oct 10 18:33:22 Tower kernel: ata8: SATA link down (SStatus 0 SControl 300)
Oct 10 20:42:52 Tower kernel: ata8: exception Emask 0x10 SAct 0x0 SErr 0x4010000 action 0xe frozen
Oct 10 20:42:52 Tower kernel: ata8: edma_err_cause=00000010 pp_flags=00000000, dev connect
Oct 10 20:42:52 Tower kernel: ata8: SError: { PHYRdyChg DevExch }
Oct 10 20:42:52 Tower kernel: ata8: hard resetting link
Oct 10 20:42:58 Tower kernel: ata8: link is slow to respond, please be patient (ready=0)
Oct 10 20:43:02 Tower kernel: ata8: SRST failed (errno=-16)
Oct 10 20:43:02 Tower kernel: ata8: hard resetting link
Oct 10 20:43:08 Tower kernel: ata8: link is slow to respond, please be patient (ready=0)
Oct 10 20:43:10 Tower kernel: ata8: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
Oct 10 20:43:10 Tower kernel: ata8.00: ATA-9: WDC WD80EFAX-68LHPN0, 7SGG4UMC, 83.H0A83, max UDMA/133
Oct 10 20:43:10 Tower kernel: ata8.00: 15628053168 sectors, multi 0: LBA48 NCQ (depth 31/32)
Oct 10 20:43:11 Tower kernel: ata8.00: configured for UDMA/133
Oct 10 20:43:11 Tower kernel: ata8: EH complete
Oct 10 20:43:11 Tower kernel: sd 11:0:0:0: [sdo] 15628053168 512-byte logical blocks: (8.00 TB/7.28 TiB)
Oct 10 20:43:11 Tower kernel: sd 11:0:0:0: [sdo] 4096-byte physical blocks
Oct 10 20:43:11 Tower kernel: sd 11:0:0:0: [sdo] Write Protect is off
Oct 10 20:43:11 Tower kernel: sd 11:0:0:0: [sdo] Mode Sense: 00 3a 00 00
Oct 10 20:43:11 Tower kernel: sd 11:0:0:0: [sdo] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA
Oct 10 20:43:11 Tower kernel: sdo: sdo1
Oct 10 20:43:11 Tower kernel: sd 11:0:0:0: [sdo] Attached SCSI disk

 

[JorgeB]

Those controllers have issues with some more recent devices, best bet would be to upgrade them.

[thenem]

just wanted to quickly comment on this thread. I also have one of the older 24 bay TAM supermicro servers with the  SAT2-MV8 controllers, I've never had a kernel panic /server crash issue like this up until these 8tb drives (for reference i have 3 6tb drives 1x WD Blue and 2x Toshiba x300 running rock solid on the latest stable build of unRAID for  ~8 months now). 

 

I picked picked up one of the WD 8tb easystores (bestb exclusive) during BF (NEBB model - White label drive inside vs the NESN model - generally red label drives inside). I've tried a few things causing in a kernel panic, preclear (plugin), opening the NERD tools plugin and searching for updates, letting the drive sit idle while the array is running.

 

I can attach my full syslog later but my kernal panic returns this - and returning a "kernel BUG at drivers/ata/stat_mv:c2118"

 

I'm planning on trying to update to the latest mainline unRAID builds later today to see if the updated Linux Kernel (which hopefully includes a newer version of the sata_mv driver) resolves my issues with these 8tb drives. 

 

I have a close friend with an identical supermicro box from TAM and he's been running multiple 8tb seagates without issue and recently installed a wd 8tb red label (he wasn't able to preclear - had crashing issues, but was able to simply add it to the arrary) and seems to not be having any of these issues. 

 

Any update on where things stand with your server?

[bobmagnuson]

I wasn't able to get past the problem with the WD Red drives.  My solution was to swap the motherboard, processors, etc. and use Dell H310's (cross flashed).  

 

Works like a champ!

[thenem]

Quick follow up for others that run across this thread as i've been PM'd a few times. 

 

My issue was 100% related to the SAT2-MV cards, i ended up ordering a LSI SAS 9201-16i pci card which covers all of my existing drives at this time - and will hopefully provide a bit of future proofing. I was able to pop in the LSI card, everything was plug n play, it detected all 8tb drives, unraid booted without issue and i was finally able to update my server.  

 

I have a friend with and identical TAM box, same SAT2 cards (that i'm aware of) and he for some reason hasn't run into a single issue with his hardware with 8tb drives (shucked). 

 

I cant seem to find any answer for why this happens other than what has been documented around the forum stating that the marvell cards + nvidia chipsets just don't play nice with unRAID.

 

Good Luck!

[plahiry]

Just a quick addition in case anyone is looking for a short-term solution - I have a TAM box and I ran into this issue with a shucked 8tb drive. I opted to just disconnect the slot in which the 8tb drive was in from the SAT2 card and hook it up to one of the SATA ports on the motherboard itself. 8tb drive booted up fine and currently pre-clearing.

 

This will give some time to source the right parts.

 

Cheers