atom c series

[l@e]

I'm thinking to make a new nas for the office. currently still using 7 and ufs.
I was thinking of Intel® Atom processor C2758, SoC, on A1SRi-2758F supermicro board.
is atom powerful enough for using raid 10 in zfs? I saw that this cpu has AES crypto instruction, can this be used on freenas? somebody knows any other board working with the same cpu that is reliable for having 24/7 ops? thanks for the advise.

 

[enemy85]

Unfortunately the new atoms seem not to be suitable for freenas :(

 

[l@e]

Unfortunately the new atoms seem not to be suitable for freenas :(

thanks for the share but what is the problem with atom cpu? or maybe the board is not suitable? I saw an old post on the forum that was recommending the micro server of HP n40l/n54l. that one is on atom.

 

[enemy85]

http://forums.freenas.org/index.php?threads/STH's-take-on-the-new-Avotons.15543/

 

[l@e]

http://forums.freenas.org/index.php?threads/STH's-take-on-the-new-Avotons.15543/

I see. just saw that the hp micro server is not with atom. :(

 

[enemy85]

well let's hope a future version of freenas will be suitable for that board too...it's a great board that!

 

[jgreco]

Unfortunately the new atoms seem not to be suitable for freenas :(

This may be a bit of a stretch. It actually booted the kernel and then fell to an NMI, and I don't think Mr. Kennedy did much to research that. Just as with USB3 support, finding the right BIOS options can be a bit tricky for FreeBSD early adopters, and finding the right defaults is something manufacturers often get a bit wrong.

It would be better to say that we are still awaiting success stories. I for one think it looks very promising.

 

[enemy85]

It would be better to say that we are still awaiting success stories. I for one think it looks very promising.

well of course with a full testing we would be more aware of the real possibilities of this board...but i guess that most of the people won't buy it without being sure of a good percentage of success...so the question is: who is gonna be the first to try? ;)

 

[jgreco]

well of course with a full testing we would be more aware of the real possibilities of this board...but i guess that most of the people won't buy it without being sure of a good percentage of success...so the question is: who is gonna be the first to try? ;)

If I didn't have so much cr*p on my plate already, it'd probably be me. But I picked up a new heavy contract and most of my free time has vanished. I am very interested in trying one of those u-nas chassis with a mini-ITX.

 

[enemy85]

If i had the money it would have been me! :)

 

[l@e]

This may be a bit of a stretch. It actually booted the kernel and then fell to an NMI, and I don't think Mr. Kennedy did much to research that. Just as with USB3 support, finding the right BIOS options can be a bit tricky for FreeBSD early adopters, and finding the right defaults is something manufacturers often get a bit wrong.

It would be better to say that we are still awaiting success stories. I for one think it looks very promising.

I agree that is strange problem. and you are right about we are waiting for somebody to test it. when first release of freenas 8 I had in hand a sm d525 with 4g ram
that I had to deliver to a project that time, I tried to install the new fn with zfs support and it worket fine. I did not run lot of test on it because of the limited time, but throughput was not bad at all. using 2 sata WD RE4 1TB in mirror. cpu was most of the time less than 25%. Bottom line I don't think that atom is not suitable for small application. this board was with 6 stata ports.

 

[l@e]

If i had the money it would have been me! :)

same status: no money >no fun :(

 

[cyberjock]

I'd definitely be willing to give it a try if I had the money. For those that know me I definitely have the time ;)

 

[jonnn]

I fall into the camp that I want some kind of assurance of a good chance of success.
I have the money, a Lian Li PC-25 sitting here begging to be used and 6 hard drives sprawled about waiting for a motherboard.

But I know little about FreeBSD, however perhaps I could be in if some of you knowledgeable about guys could offer some help with these "NMI" issues? (I don't even know what an NMI is...)

I know the board has "IPMI" would it be possible to give access over the internet so the system can be troubleshooted/worked on remotely? I do have a router with DD-WRT and a fast connection if that helps (that's about my experience with "alternative" operating systems.)

I am looking at the supermicro board or the asrock board.

 

[DrKK]

But I know little about FreeBSD, however perhaps I could be in if some of you knowledgeable about guys could offer some help with these "NMI" issues? (I don't even know what an NMI is...)

An NMI is a pretty bad show-stopper problem. Here is the wikipedia on it.

By the way, I have a new build, and I'm using the Supermicro X10-SLM+-F-O board with the G3220 Intel Haswell, and it's totally rocking out. I'm becoming a big fan of the Supermicro stuff now.

 

[cyberjock]

It's like DrKK said. It's a pretty serious issue. Usually that's a hardware error that's thrown. But software that isn't written for the hardware can cause errors to be thrown too. More than likely its going to take some developers some time to get it to work. My guess is that the FreeBSD developers will get it working and then it'll be rolled into FreeNAS at some point.

 

[gpsguy]

Maybe iXsystems could be persuaded to look at it.

Might be the basis for a new superMini with ECC RAM support.:)

 

[cyberjock]

My guess is the iX guys have more important things to slay and figure the FreeBSD guys should deal with it instead of them.

 

[Dusan]

Unfortunately the new atoms seem not to be suitable for freenas :(

http://forums.freenas.org/index.php?threads/STH's-take-on-the-new-Avotons.15543/

STH tested an preproduction ASRock board. An NMI is more a board issue than a CPU one. I checked the Supermicro A1SRi-2758F manual and it has the usual SM NMI stuff (the same as my X9SCL) -- a hardware NMI button header (don't connect anything here) and a Watch Dog Enable jumper (set to Disable). So, it's very well possible that the Supermicro Avoton boards work, just nobody tried it yet.

 

[jonnn]

STH tested an preproduction ASRock board. An NMI is more a board issue than a CPU one. I checked the Supermicro A1SRi-2758F manual and it has the usual SM NMI stuff (the same as my X9SCL) -- a hardware NMI button header (don't connect anything here) and a Watch Dog Enable jumper (set to Disable). So, it's very well possible that the Supermicro Avoton boards work, just nobody tried it yet.

I think STH is the only place on the planet with a A1SRi-2758F. And their reviews leave a lot to be desired.
Sigh.
What would happen if I sent a A1SRi-2758F or Asrock board to the iXsystems team? :-D