SuperMicro X10SL7-F

[grep137]

Ziferius, I'm getting some strange results from one of the modules I got from Superbiiz back in March (see below).

I ran Memtest86+ 5.01 for 17 hours on the default settings and it found no errors. However, when I looked at the Event Logs in the BIOS it showed (see below). I've only had the machine powered on fewer than 10 times (on for only 1 - 5 hours each time, except for the 17 hour memtest) since I put it together. Is this normal, or is DIMMA1 faulty?

DATE: | TIME: | ERROR CODE: | SEVERITY: | DESCRIPTION:
05/05/14 | 01:20:30 | Smbios 0x01 | DIMMA1 | Single Bit ECC Memory Error
05/05/14 | 03:08:27 | Smbios 0x01 | DIMMA1 | Single Bit ECC Memory Error
05/05/14 | 06:48:06 | Smbios 0x01 | DIMMA1 | Single Bit ECC Memory Error
05/05/14 | 09:34:40 | Smbios 0x01 | DIMMA1 | Single Bit ECC Memory Error
05/18/14 | 11:48:03 | Smbios 0x01 | DIMMA1 | Single Bit ECC Memory Error
06/15/14 | 11:32:58 | Smbios 0x01 | DIMMA1 | Single Bit ECC Memory Error
06/22/14 | 15:21:03 | Smbios 0x01 | DIMMA1 | Single Bit ECC Memory Error

Samsung DDR3-1600 ECC CL11
Mfr Part Number: M391B1G73QH0-YK0

 

[Sir.Robin]

Where are you getting your memory, if you don't mind me asking?

Thanks!

Sent from my iPad using Tapatalk

I have Kingston for my primary system. My secondary will run on Crucial ECC initially. Might exchange that later.

Have you checked CPU then? Bent pins?

 

[Sir.Robin]

Thank you for the firmware flashing instructions Sir.Robin.

You're welcome! :)

 

[Ziferius]

Ziferius, I'm getting some strange results from one of the modules I got from Superbiiz back in March (see below).

I ran Memtest86+ 5.01 for 17 hours on the default settings and it found no errors. However, when I looked at the Event Logs in the BIOS it showed (see below). I've only had the machine powered on fewer than 10 times (on for only 1 - 5 hours each time, except for the 17 hour memtest) since I put it together. Is this normal, or is DIMMA1 faulty?

DATE: | TIME: | ERROR CODE: | SEVERITY: | DESCRIPTION:
05/05/14 | 01:20:30 | Smbios 0x01 | DIMMA1 | Single Bit ECC Memory Error
05/05/14 | 03:08:27 | Smbios 0x01 | DIMMA1 | Single Bit ECC Memory Error
05/05/14 | 06:48:06 | Smbios 0x01 | DIMMA1 | Single Bit ECC Memory Error
05/05/14 | 09:34:40 | Smbios 0x01 | DIMMA1 | Single Bit ECC Memory Error
05/18/14 | 11:48:03 | Smbios 0x01 | DIMMA1 | Single Bit ECC Memory Error
06/15/14 | 11:32:58 | Smbios 0x01 | DIMMA1 | Single Bit ECC Memory Error
06/22/14 | 15:21:03 | Smbios 0x01 | DIMMA1 | Single Bit ECC Memory Error

Samsung DDR3-1600 ECC CL11
Mfr Part Number: M391B1G73QH0-YK0

Interesting... That seems like an aweful lot, but that's why we run ECC....it catches and fixes that.
Cyber jock is definitely proved right (yet again) why ECC is so important.




Sent from my iPad using Tapatalk

 

[Ziferius]

I have Kingston for my primary system. My secondary will run on Crucial ECC initially. Might exchange that later.

Have you checked CPU then? Bent pins?

CPU socket on first board had bent pins... I had a pentium G3220 for it...


Sent from my iPad using Tapatalk

 

[grep137]

I'm glad I listened and got ECC. It does seem a little strange that all the errors are occurring on only one DIMM (that is if I'm reading the event log correctly). Good luck getting your board to work -- sorry I can't offer any help.

 

[Ziferius]

It could be the DIMM or it could be the socket. To check though - trade sockets with that DIMM with one of the others. If the problem follows - then it's the DIMM. If it doesn't, then it's the socket.

 

[grep137]

It could be the DIMM or it could be the socket. To check though - trade sockets with that DIMM with one of the others. If the problem follows - then it's the DIMM. If it doesn't, then it's the socket.

Thanks. I'll switch them and see what happens. Hopefully it's not the board.

 

[Sir.Robin]

Thanks. I'll switch them and see what happens. Hopefully it's not the board.

I would say it's more likely a bad dimm (memory stick).

 

[Sir.Robin]

Awh crap!! They slashed the pricing of Samsung 8GB M391B1G73BH0-CK0 modules!! Just a tiny bit more than what i gave for the kingston's

 

[grep137]

Ran Memtest at the default settings again for another 18 hours -- still no errors. I wonder if the errors are corrected by the ECC before Memtest can detect them. I was also thinking, since Memtest can read and write to the DIMM with no errors, would this definitely absolve the motherboard?

I would say it's more likely a bad dimm (memory stick).

Thanks Sir.Robin. Also, if you get the Samsungs just don't sit on them for a few months, like I did (see below).

Ziferius, I thought I'd get the ball rolling just in case it takes a few days to get a response... Don't know if you've tried contacting Samsung yet about your memory, however, here is the response I received:
Unfortunately you've reached the Retail Memory department. The part number
you've provided is for a OEM unit meant to be sold together with Pre-built
systems (Dell, HP, ETC). You can try contacting your point of purchase or
system manufacturer for further assistance.

Contacted the system manufacturer and he's pretty much clueless too.

Looks like getting a replacement from Samsung may be a little tricky.

 

[Ericloewe]

Ran Memtest at the default settings again for another 18 hours -- still no errors. I wonder if the errors are corrected by the ECC before Memtest can detect them. I was also thinking, since Memtest can read and write to the DIMM with no errors, would this definitely absolve the motherboard?

There are tales that Memtest 86+ detects ECC-corrected (and detected) errors and logs them... I say tales because I've never seen proof and the whole figuring out if ECC is working business relies on blind faith that ECC is working, in many cases.

 

[grep137]

There are tales that Memtest 86+ detects ECC-corrected (and detected) errors and logs them... I say tales because I've never seen proof and the whole figuring out if ECC is working business relies on blind faith that ECC is working, in many cases.

Ericloewe, thank you for enlightening (and frightening :) ) me.

Ziferius, just got a response from Superbiiz and it looks like they will let me RMA the module -- I can post more info if you'd like.

 

[Ziferius]

cool. Grep137 - has it been more than 30 days?
I went ahead and requested an RMA as well.

I've tried everything that I could have. I did order 32gb of Hynix last night ... we'll see if this works :)
I'm really looking forward to getting this system going!. (Though, I have to get 3 more 3 TB drives. . . .)

 

[Z300M]

Awh crap!! They slashed the pricing of Samsung 8GB M391B1G73BH0-CK0 modules!! Just a tiny bit more than what i gave for the kingston's

Am I correct in assuming that the "they" who slashed the price on the Samsung modules is a European vendor? I don't see a price drop by any of the US vendors.

 

[Z300M]

cool. Grep137 - has it been more than 30 days?
I went ahead and requested an RMA as well.

I've tried everything that I could have. I did order 32gb of Hynix last night ... we'll see if this works :)
I'm really looking forward to getting this system going!. (Though, I have to get 3 more 3 TB drives. . . .)

Where did you find your Hynix modules, and how much?

 

[Sir.Robin]

Am I correct in assuming that the "they" who slashed the price on the Samsung modules is a European vendor? I don't see a price drop by any of the US vendors.

Hehe... yeah. Norwegian/danish shop. The samsung sticks were cinda pricey before... but now that they are over 30% lower it's abut the same price as i payed for the kingston sticks.

 

[grep137]

cool. Grep137 - has it been more than 30 days?
I went ahead and requested an RMA as well.

I got mine back in March. Superbiiz's response when I told them the purchase date: "Please note that because you are beyond the 30 day period, you are no longer eligible for a refund, only repair/replacement." However, this response wasn't from someone who actually works in the RMA dept.

Just found this too:
"Items received by SuperBiiz after the initial 30-day period are no longer eligible for refund and will be forwarded to the original manufacturer for repair or replacement only, the choice of which will be at the discretion of the manufacturer."
http://www.superbiiz.com/rma.php

I believe that they may still do replacements beyond the 30 days because of Samsung's policy on ECC memory (copied from the e-mail I received from Samsung): "You can try contacting your point of purchase or system manufacturer for further assistance."

Which makes me wonder who we would contact in the future if Superbiiz goes out of business.

Also, I tried to get them to cross ship, but that seems like a no go so far; maybe you'll have better luck.

 

[Kuro Houou]

I just got my X10Sl7F mobo and have everything setup, updated all the firmware's and installed the LSI IT Mode Firmware from here, ftp://ftp.supermicro.com/Driver/SAS/LSI/2308/Firmware/IT/ version 19... I can't see my drives though, they aren't showing up in the SAS configuration. Was version 19 the right one to use?

 

[Sir.Robin]

I just got my X10Sl7F mobo and have everything setup, updated all the firmware's and installed the LSI IT Mode Firmware from here, ftp://ftp.supermicro.com/Driver/SAS/LSI/2308/Firmware/IT/ version 19... I can't see my drives though, they aren't showing up in the SAS configuration. Was version 19 the right one to use?

Congrats on a great motherboard :)

No, FreeNAS is on the ver 16 driver level. And so your firmware also needs to be there.