ECC Error - Problem?

[TMR812]

I noticed, while checking my memory status apple:, about this mac, more info, memory), that one of my memory stick had an ECC error. Specifically, it had one error that was correctable.

I've never seen this before, and haven't since. This stick is part of new recently purchased memory from OWC. I've run 4 or 5 passes of testing using "Rember", and have found no errors. Since the error, I ran 2 passes of "Rember", and found no errors, and haven't seen one for several days since.

Question: Do I have a problem, or is the occasional "correctable" error normal with ECC memory? Should I be looking to RMA this stick right away, or just keep an eye on it?

Thanks for the input,

TMR

 

[HW-Hack]

Well thats why you have ECC memory - so you can correct either the occasional bit-flip cause by a soft-memory error or a raging alpha-particle.

http://www.siliconfareast.com/soft-error.htm

The rest of us without ECC just have to make due and keep our heads low

 

[CanadaRAM]

The Apple standard is that 4 Correctable ECC errors in 48 hours is normal.

 

[TMR812]

Thanks

Thanks for the responses. Sounds like you both think this is perfectly normal.

Thanks,

TMR

 

[jczubach]

The Apple standard is that 4 Correctable ECC errors in 48 hours is normal.

next time you're in vancouver, look me up, you rascal. I will definitely foot for the pitcher at the pub of your choice.

 

[Anonymous Freak]

Thanks for the responses. Sounds like you both think this is perfectly normal.

Thanks,

TMR

Yup. If you were on a non-ECC computer, you probably would have just had an unexplained crash. (Possibly just a program, possibly the whole OS, impossible to tell.) It's also possible that it could have been in an area that wasn't active, so the error might have just gone unnoticed. That's the joy of ECC, it doesn't matter. A random memory error every once in a while (especially if it passes stress tests just fine,) is reasonably normal. I've seen servers with perfectly good RAM that have been in service for 4+ years with a few dozen ECC errors reports over the years. As long as they don't start getting very regular, or decidedly repeatable, you're fine. As caRAM points out, Apple considers 4 in a 48 hour span to be just fine. That means up to 60 a month would be acceptable. I wouldn't consider 60 in a month to be acceptable, but as long as it passes stress tests, I suppose it could be considered 'okay'.

 

[PMG5Quad]

Mac Pro ECC doing what it should or not?

Yup. If you were on a non-ECC computer, you probably would have just had an unexplained crash. (Possibly just a program, possibly the whole OS, impossible to tell.) It's also possible that it could have been in an area that wasn't active, so the error might have just gone unnoticed. That's the joy of ECC, it doesn't matter. A random memory error every once in a while (especially if it passes stress tests just fine,) is reasonably normal. I've seen servers with perfectly good RAM that have been in service for 4+ years with a few dozen ECC errors reports over the years. As long as they don't start getting very regular, or decidedly repeatable, you're fine. As caRAM points out, Apple considers 4 in a 48 hour span to be just fine. That means up to 60 a month would be acceptable. I wouldn't consider 60 in a month to be acceptable, but as long as it passes stress tests, I suppose it could be considered 'okay'.

Just so I understand your saying my ECC ram is helping me (it makes sense that is the reason we payed for a Mac Pro less problems for us the end user)

Please tell me what you think of this it has been this way 35 Hours
I bought this ram from DMS a week ago. I am not saying anything but is this something I need to replace or is it good ram doing a great job?

DIMM Riser B/DIMM 4:

Size: 2 GB
Type: DDR2 FB-DIMM
Speed: 667 MHz
Status: ECC Errors
ECC Correctable Errors: 761744
Manufacturer: 0x0000
Part Number: 0x000000463732353642363145353636374600
Serial Number: 0x00000000

Thank You,
Thomas

 

[canonballs]

The rest of us without ECC just have to make due and keep our heads low

That's the thing with ECC—the manufacturers can have an occassional error and still stay within spec. Effectively this allows them to relax QC standards. Being cynical I suspect they use this circumstance to their best advantage.

 

[Anonymous Freak]

Please tell me what you think of this it has been this way 35 Hours
I bought this ram from DMS a week ago. I am not saying anything but is this something I need to replace or is it good ram doing a great job?

ECC Correctable Errors: 761744

That is a bit extreme. Even one correctable error a day would be likely 'acceptable', but that many in one week is beyond ridiculous.

Make sure the module is seated properly, and move it to a different socket to ensure that the errors follow the module, not the socket.

If you continue to get that many errors in such a short time span; return the memory, it's bad.

 

[xraydoc]

I get ECC errors like that if I let my Mac Pro run at standard fan RPMs. If I use the smc fan applet to boost my Mac's fan speeds to 1100-1200 rpm, the errors stop completely. Fan noise only goes up a little.

This, however, is with 4 hard drives, all RAM slots occupied and two video cards.

 

[pprior]

Where do you see the ECC report?

I have a 2008 mac pro and under memory I see 8 memory slots, size 2gb, type DDR2-FB-DIMM and speed 800Mhz, Status says OK.

Are you seeing ECC error under status? or is there somewhere else to look?

Just wondering.

 

[ronstone]

Please tell me what you think of this it has been this way 35 Hours
I bought this ram from DMS a week ago. I am not saying anything but is this something I need to replace or is it good ram doing a great job?

DIMM Riser B/DIMM 4:

Size: 2 GB
Status: ECC Errors
ECC Correctable Errors: 761744

Hi:
What did you decide to do? Have you returned the system? Did the errors go away?
One thing to check is to switch the DIMMs around in the slots. e.g. rotate each one into a new position.
a) if the errors follow the DIMM to a new position, then return the DIMM.
b) if the errors are reported in the same slot. Return the MAC
c) if the errors go away, you're happy.

 

[jimgeb]

Bad Mac Pro

I wanted everyone to know who has a ecc error problem on their new 2009 Mac Pro (2.66 quad) that I had to have my first one replaced by Apple because I kept getting ecc correctable errors in the 3rd and 4th memory slots while doing some heavy rendering in Adobe After Effects CS4. I switched the memory modules around to be sure it wasn't the memory. The errors only appeared in the 3rd and 4th slots. I'm on my second Mac Pro now and everything seems OK. Apple was great about it and replaced it once I sent them a copy of my system profile which showed the errors. Both the Apple extended hardware test and memtest showed no problems. It only showed up when the memory was being loaded up with rendering tasks.