Jump to content

Looking for a specific motherboard


eyeball

Recommended Posts

Hi Everyone,

I dont know where to look to narrow my search so if anyone could help that would be fantastic. Basically I am looking for a motherboard with the following qualities:

support socket Intel 775 (Q6600 to be exact)

Support DDR2 (must be ECC RAM)

have 2 PCI x16 slots

have 2 PCI-express 1x (the really small ones)

that is all really, if anyone knows of a decent one i would appreciate your help, otherwise I will keep on looking :)

Thanks

Link to comment
Share on other sites


Why does it need to be ECC RAM? You know you will need a server/workstation chipset for that.

^^ What he said.

There's not too many chipsets that fit the bill (Socket 775 + ECC). There's the 5 year old 975X which is kinda crappy. And then X38 & X48 which are quite hard to find now and have mostly moved to DDR3 (and often only detect ECC errors/doesn't correct them -- and that only with DDR2), as well as 3200/3210 which in most cases will have PCI-X slots (or PCI-e x8/x4) instead of PCI-e x16 (some of those would use FB-DIMMs instead too).

The closest thing I can find right now is the $240 TYAN S5211G2NR Toledo i3210W. Socket 775, 3210/ICH9R chipset, 8GB DDR2 max (4 slots, ECC supported), etc. Your two x16 slots will work at x8 though as there's only 16 PCI-e lanes total.

Not that I would bother with ECC RAM on a home server. Not quite sure why you'd want two PCI-e x16 slots for that either.

Link to comment
Share on other sites

Indeed, ditch the ECC-RAM idea. I think I know why you brought it up as older times told you to use ECC-RAM in a Server, but that was 10 years ago so say welcome to the modern world we live in now, hell, not even google is using ECC-RAM in their servers ;).

Now, the 2 PCI-E 16x (electricaly?) sloths is indeed a good question from coffeefiend.

So, you have a Q6600 you want to re-use? Get a P45 based mobo (GIGABYTE GA-EP45T-USB3P is a great deal at some 130USD) for that with DDR3.

Link to comment
Share on other sites

Indeed, ditch the ECC-RAM idea. I think I know why you brought it up as older times told you to use ECC-RAM in a Server, but that was 10 years ago so say welcome to the modern world we live in now, hell, not even google is using ECC-RAM in their servers ;).

Now, the 2 PCI-E 16x (electricaly?) sloths is indeed a good question from coffeefiend.

So, you have a Q6600 you want to re-use? Get a P45 based mobo (GIGABYTE GA-EP45T-USB3P is a great deal at some 130USD) for that with DDR3.

Is ECC Ram not used as often now? :blink:

What happens with in memory errors? flipped bits and all that? dont they make it down to disk and corrupt the filesystem anymore?

The 2 PCI x16 slots are because I have a raid controller to pop in there too and Im thinking If I stick it into the only x16 slot it might not work, at least with 2 i could have a low end vga card and the raid controller in the other slot and it should detect both ok, or maybe Im wrong about this one but I just wanted to make sure it would work.

and both those boards look great thank you but I really want/need ECC Ram......do you know of any whitepapers on ECC vs non ECC I could refer too? ... im still not comfortable throwing non ECC at it. :)

Thanks again

Edited by eyeball
Link to comment
Share on other sites

Is ECC Ram not used as often now? :blink:

What happens with in memory errors? flipped bits and all that? dont they make it down to disk and corrupt the filesystem anymore?

The same happens as on your regular PC: usually nothing, or very close to that. My question is: and what do you think happens when a CPU does something similar (they all have errata)? Or when a network packet is corrupted? Or when a bit is corrupted on magnetic storage? Or a driver (or the OS) that has a bug screws up some data somehow? (all of these being FAR more common) It's a massive waste of money IMO. Even more so on a home server -- it's not like you're providing services that can't ever go down, or that you are going to countless thousands of dollars every time it does.

The 2 PCI x16 slots are because I have a raid controller

Most fancy RAID controllers don't even require a x16 slot. I have some (the cheapo kind admittedly) in x1 slots. I think the most I've seen is a x8, but that's serious overkill. Sure, you can saturate a simple x1 card in RAID as it's limited to 250MB/s (that's still more than acceptable performance in my book for a home server though), but even just a card that fits in a x4 slot has access to 1GB/s of bandwidth -- you're going to need a lot of really fast hard drives to saturate this, or preferably RAID'ing expensive SSDs! A PCI-e x16 2.0 slot has a 8GB/s BW which you probably wouldn't hit, even with 16 high end SSDs in RAID so it doesn't make much sense for them to use that. Then again, your network is nowhere near that fast (good Gigabit ethernet stuff with jumbo frames and everything has about half the BW of a single x1 slot -- that's ultimately how fast you'll be able to get data off from it). Just sit down for a sec and think what you're going to need this for. Personally, I hardly ever need very large files really fast. Most large files I access are movies and the speed required for this is pathetic (around 1MB/s -- a ATA33 drive from the 90's could easily handle it speed-wise)

Link to comment
Share on other sites

Is ECC Ram not used as often now? :blink:

What happens with in memory errors? flipped bits and all that? dont they make it down to disk and corrupt the filesystem anymore?

The same happens as on your regular PC: usually nothing, or very close to that. My question is: and what do you think happens when a CPU does something similar (they all have errata)? Or when a network packet is corrupted? Or when a bit is corrupted on magnetic storage? Or a driver that has a bug screws up some data somehow? (all of these being FAR more common) It's a massive waste of money IMO. Even more so on a home server -- it's not like you're providing services that can't ever go down, or that you are going to countless thousands of dollars every time it does.

The 2 PCI x16 slots are because I have a raid controller

Most fancy RAID controllers don't even require a x16 slot. I have some (the cheapo kind admittedly) in x1 slots. I think the most I've seen is a x8, but that's serious overkill. Sure, you can saturate a simple x1 card in RAID as it's limited to 250MB/s (that's still more than acceptable performance in my book for a home server though), but even just a card that fits in a x4 slot has access to 1GB/s of bandwidth -- you're going to need a lot of really fast hard drives to saturate this, or preferably RAID'ing expensive SSDs! A PCI-e x16 2.0 slot has a 8GB/s BW which you probably wouldn't hit, even with 16 high end SSDs in RAID so it doesn't make much sense for them to use that. Then again, your network is nowhere near that fast (good Gigabit ethernet stuff with jumbo frames and everything has about half the BW of a single x1 slot -- that's ultimately how fast you'll be able to get data off from it). Just sit down for a sec and think what you're going to need this for. Personally, I hardly ever need very large files really fast. Most large files I access are movies and the speed required for this is pathetic (around 1MB/s -- a ATA33 drive from the 90's could easily handle it speed-wise)

Hmmmm, very good reply CoffeeFiend I shall re-consider this entire build.

Thank you for the help; as usual your knowlegde amazes me :wub:

Link to comment
Share on other sites

So, what parts do you have already besides the Q6600?

Ok you asked for it.... here is the whole story.

I built a server at the back end of 2008 that has the following:

Areca ARC-1230 Card with 1GB of ECC Ram.

2GB of DDR2 Ram

Intel Q6600

An Asus Desktop Motherboard (I forget the model but its just a standard desktop board)

12 x 1TB Drives.

All was good until I noticed I was getting errors every now and again in the event log, NTFS ID55 errors IIRC. Im running server 2008 and self healing NTFS kicks in and repairs the problem and everything is fine again.

But it is annoying me like mad, I just want things to run without these errors. I have worked with countless servers and never seen an error as frequent as this, So I am lookinng at replacing the ram + board for something ECC Compatible to see if it stops the errors.

Thanks

Link to comment
Share on other sites

It's rather unlikely to be a RAM problem, but I'd still run memtest86+ to check it anyway. Ideally you'd check the RAID card's RAM too but that's a little trickier.

First thing I'd do is check the SATA cables (and make sure they are the locking kind), and making sure I have the latest firmware on the RAID card along with the latest drivers for it (and Windows fully patched of course). There's also still the possibility of having one or more bad sectors on any of your drives (don't rely on SMART to let you know). It also could be caused by any piece of software that interferes with disk I/O, like an antivirus (or even malware).

There's also the possibility that the error ID 55 isn't actually a "real" problem as seen here. It would also help knowing more about the particular problem, like if there are any other errors logged around the same time.

There's a lot of possibilities which are FAR more likely than "my RAM flips bits randomly a billion times more often than it does in any other computer", and I'd say the odds ECC would fix any of this are right next to none (I'd bet good money it won't fix it)

Link to comment
Share on other sites

It's rather unlikely to be a RAM problem, but I'd still run memtest86+ to check it anyway. Ideally you'd check the RAID card's RAM too but that's a little trickier.

First thing I'd do is check the SATA cables (and make sure they are the locking kind), and making sure I have the latest firmware on the RAID card along with the latest drivers for it (and Windows fully patched of course). There's also still the possibility of having one or more bad sectors on any of your drives (don't rely on SMART to let you know). It also could be caused by any piece of software that interferes with disk I/O, like an antivirus (or even malware).

There's also the possibility that the error ID 55 isn't actually a "real" problem as seen here. It would also help knowing more about the particular problem, like if there are any other errors logged around the same time.

There's a lot of possibilities which are FAR more likely than "my RAM flips bits randomly a billion times more often than it does in any other computer", and I'd say the odds ECC would fix any of this are right next to none (I'd bet good money it won't fix it)

I have run Memtest 86+ for 12 hours on the server and found nothing. The Card originally had 256mb ram and still gave me the errors, I upgraded to 1gb in the hope of eliminating the issue but I still get the errors So I think I can say the Card's Ram is ok.

I am considering a reinstall anyway so now might the time. With regards to drivers for the card I have a choice of Storport or Scsiport. Currently I am using Scsiport at the moment as I remember storport bluescreens when i/O becomes heavy. I think a reinstall might be good for it, at least then I can try to use Storport drivers again, who knows they may have updated them since 2008.

Also I will look into buying locking sata cables, I wasnt aware they even existed to be honest And I will certainly have a thorough read through that thread you linked to.

Thanks again for the help :)

Link to comment
Share on other sites

  • 1 month later...

Sorry to drag up a dead topic but I wanted to tell you that after a reinstall with Server 2008 R2, the newest Storport drivers and a firmware upgrade on the card all is well and has been for weeks :)

A big thank you to everyone who helped me with this :thumbup

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...