March 12, 201214 yr ok still flexing my way thru my first server, other day had an issue where the drives were disconnecting (see thread "is this something to be concerned about") and it was narrowed down to either a sata cable issue or power issue. replaces the sata cables, and the other day the server just ... went away. Web access wouldn't respond, console wouldn't respond.. only thing I could do was hit the reset button. Server restarted and came up, started doing a parity check (6 hours later, was fine). then in the log I noticed... Mar 9 22:02:31 Burbank kernel: ata2: hard resetting link Mar 9 22:02:31 Burbank kernel: ata6: hard resetting link Mar 9 22:02:31 Burbank kernel: ata5: hard resetting link Mar 9 22:02:31 Burbank kernel: ata1.00: irq_stat 0x08000000, interface fatal error Mar 9 22:02:31 Burbank kernel: ata1: SError: { UnrecovData HostInt 10B8B BadCRC } Mar 9 22:02:31 Burbank kernel: ata1.00: failed command: READ DMA EXT Mar 9 22:02:31 Burbank kernel: ata1.00: cmd 25/00:18:77:bf:da/00:03:15:00:00/e0 tag 0 dma 405504 in Mar 9 22:02:31 Burbank kernel: res 50/00:00:76:bf:da/00:00:15:00:00/e5 Emask 0x50 (ATA bus error) Mar 9 22:02:31 Burbank kernel: ata1.00: status: { DRDY } Mar 9 22:02:31 Burbank kernel: ata1: hard resetting link Mar 9 22:02:31 Burbank kernel: ata2: SATA link up 3.0 Gbps (SStatus 123 SControl 300) Mar 9 22:02:31 Burbank kernel: ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300) Mar 9 22:02:31 Burbank kernel: ata6: SATA link up 3.0 Gbps (SStatus 123 SControl 300) Mar 9 22:02:31 Burbank kernel: ata5: SATA link up 3.0 Gbps (SStatus 123 SControl 300) Mar 9 22:02:31 Burbank kernel: ata1.00: configured for UDMA/133 Mar 9 22:02:31 Burbank kernel: ata2.00: configured for UDMA/133 Mar 9 22:02:31 Burbank kernel: ata5.00: configured for UDMA/133 Mar 9 22:02:31 Burbank kernel: ata6.00: configured for UDMA/133 Mar 9 22:02:31 Burbank kernel: ata2: EH complete Mar 9 22:02:31 Burbank kernel: ata5: EH complete Mar 9 22:02:31 Burbank kernel: ata6: EH complete Mar 9 22:02:31 Burbank kernel: ata1: EH complete (DON'T Ask for the rest, it's gone.. don't ask) the SMART for ata1 (parity) now shows ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE 1 Raw_Read_Error_Rate 0x000b 100 100 016 Pre-fail Always - 0 2 Throughput_Performance 0x0005 137 137 054 Pre-fail Offline - 90 3 Spin_Up_Time 0x0007 150 150 024 Pre-fail Always - 331 (Average 402) 4 Start_Stop_Count 0x0012 100 100 000 Old_age Always - 102 5 Reallocated_Sector_Ct 0x0033 100 100 005 Pre-fail Always - 0 7 Seek_Error_Rate 0x000b 100 100 067 Pre-fail Always - 0 8 Seek_Time_Performance 0x0005 144 144 020 Pre-fail Offline - 30 9 Power_On_Hours 0x0012 100 100 000 Old_age Always - 667 10 Spin_Retry_Count 0x0013 100 100 060 Pre-fail Always - 0 12 Power_Cycle_Count 0x0032 100 100 000 Old_age Always - 33 192 Power-Off_Retract_Count 0x0032 100 100 000 Old_age Always - 102 193 Load_Cycle_Count 0x0012 100 100 000 Old_age Always - 102 194 Temperature_Celsius 0x0002 230 230 000 Old_age Always - 26 (Lifetime Min/Max 19/46) 196 Reallocated_Event_Count 0x0032 100 100 000 Old_age Always - 0 197 Current_Pending_Sector 0x0022 100 100 000 Old_age Always - 0 198 Offline_Uncorrectable 0x0008 100 100 000 Old_age Offline - 0 [color=red]199 UDMA_CRC_Error_Count 0x000a 200 200 000 Old_age Always - 1[/color] all others are clean (I have done some research and general consensus says this isn't a huge thing) ok now as for the 'bigger boat' Right now I have a Cooler Master Silent Pro M 700W in there... single 12v rail, says 50a available drives are Hitachi 5k3000 2tb, which from what I can see draw 2a at idle (spun down), 5a idle (spinning) and 7a seek. I have 6 of them so assuming that's right... with 6 drives spinning, I got a 30a draw. out of 50. If I start doing a lot of access (XBMC updating, copying new data)... I could easily get seeking going on 3 or 4 of those at the same time.. lets say 3 are seeking... that means I'm up to 3x5a + 3x7a .. 36a out of 50. nevermind what he rest of the system is using.. and I have 4 more drives to put in.. also Hitachi 5k3000 2tb... wattage I'm fine... power on that 12v rail.... I think I'm riding the edge right now... and may have fallen over a couple times. I think I need a bigger boat? thoughts?
March 12, 201214 yr Author I've changed it twice I'll change it again.. however, I am concerned on the power issue on the 12v rail. is 50a enough to keep 10 of the Hitachi 5K3000's spinning? and run the rest of the server? Note that FOUR drives were reset... ata1, ata2, ata5 and ata6... so power or 4 bad cables?
March 12, 201214 yr If its not the SATA cable then it is a power connection problem. Do the drives have any connections in common?
March 12, 201214 yr Author the powersupply has 3 sata connecters on each cable, and those are all 4 on different cables... 1, 2-5, 3-6, 4 (just spread them out for a while to check things)
March 12, 201214 yr My thought - don't confuse units. Those numbers you quoted are watts, not amps. 7W = 0.58A.
March 12, 201214 yr Author crap on a cracker, you're right... I must have been tired when I did that.. ok I still think there's a power issue tho... it has all new cables and I've been looking at the power connectors on the CoolerMaster (flat,stranded cables, look kinda weak to me...)
March 13, 201214 yr Agreed on the 50A is fine - any good power supply will easily outperform its specs for things like disk spinup I was using a Seasonic S12II-500 which has two +12V rails, so in effect limiting my hard-drive-accessible 12V line to 17A total. I used it for two years in my UnRAID server even as I added more and more drives, never thinking about it. Before I realized I was going a bit overboard, I was running 15 2TB drives on that trusty Seasonic. I had no drive spin-up groups (all would spin up simultaneously without fail). I also performed parity checks routinely. No issues. If a new/different power supply doesn't change anything, you definitely have no power supply issue. But with a modular power supply, you have two connections to worry over: 1) the drive end, and 2) the power supply end It's possible that the factory soldering job didn't come out as intended.
March 13, 201214 yr I just wouldn't trust a CoolerMaster power supply. A few people here have had goofy drive issues with Antec supplies which had good specifications so this CoolerMaster could be the same thing.
March 13, 201214 yr Author I just wouldn't trust a CoolerMaster power supply. A few people here have had goofy drive issues with Antec supplies which had good specifications so this CoolerMaster could be the same thing. I ordered a 600w OCZ Modsteam Pro this morning.. have OCZ's in several other machines and never any issue... will report back in a few.
Archived
This topic is now archived and is closed to further replies.