Using the Supermicro X8SIL-F-O to get the handy 3x PCI-E x8, x4 slot layout I was thinking of using 3x AOC-SASLP-MV8 until I happily stumbled across a thread here that spoke of problems with the cards when bunching up three of them.

 

So, to get around that I was thinking that I d be using 2x AOC-SASLP-MV8 and for the 3rd slot, a ST LABS SATA/300 4-PORT PCI-E to get a nice and even port/disk count using 4x IB-555SSK in a PC-P80. It s 4 ports less but wth. It s not that I ve seen a case that would allow a 5th Icy Box anyway and internal drives are out of the question cus the person whos gona be using the server isnt gona be opening it up.

 

As I m a total noob doing this kind of stuff I was hoping I could get some feedback on this particular combination, if there d be any flaws I ve overlooked or anything, before I go and buy the kit.

 

cheers,

 

J

It looks like that board already has 6 sata ports on it, so combined with the 16 ports from the 2 supermicro cards will max out the Pro unRaid license anyways.... If Tom does add more drives to the license, you could always get one of the $8 ebay 2 port cards to bump it up to 24...

 

Hypothetically speaking, if you want more then 24 drives, you could try 2 MV8 boards and one of the Br10i cards. Here's a link to them on sale for $50: http://lime-technology.com/forum/index.php?topic=12177.0. They've been getting some positive feedback around here... You may want to do some reading on this forum on compatibility and shortcomings of it first though...

Yeah I'm still waiting to hear back from Supermicro as to what our options are, but bjp999 also suggested the 2xMV8+1xBr10i combo ... for $50 it's probably worth the experiment.

FYI, as it stands currently spindowns are not working with the Br10i, just be aware.

FYI, as it stands currently spindowns are not working with the Br10i, just be aware.

 

Biggest shortcoming is lack of 3T support IMO.  If LSI suddenly came out with a BIOS update to 3T I would be recommending it more.  I have one and recommend it only for people that understand the limitations.  Seems LSI is not interested is cards that grow to support larger disks in the future (guess they'd rather sell new controllers).

 

I've been using a BR10i along side an AOC-SASLP-MV8 on a MBD-X8SIL-F-O for a week now (unRAID 4.7) and it's working perfectly except for the temp and spindown.  And also I installed one of the $8 2-port cards and used it for several days without any problem.

Biggest shortcoming is lack of 3T support IMO.  If LSI suddenly came out with a BIOS update to 3T I would be recommending it more.  I have one and recommend it only for people that understand the limitations.  Seems LSI is not interested is cards that grow to support larger disks in the future (guess they'd rather sell new controllers).

 

The curious thing with LSI1068E is that according to Hitachi Ultrastar Compatibility guide it does list support for 3TB on Dell 6/iR (and this adapter is LSI1068E based - it just uses different fanout cables):

 

http://www.hitachigst.com/tech/techlib.nsf/products/Ultrastar_7K3000

My thoughts are ONE of the selling points for unRAID versus other products out there (freenas, zfs, etc...) is no Spin-Down's. Lime needs to work on getting these SAS/SATA cards working (fully) just as they do with on board SATA ports. Otherwise once you lose a selling point here and there, then people move to other solutions.

I'm not sure I fully agree with you.  With unRAID's current limit of 22 drives, two SASLP cards can be used in addition to motherboard ports to reach that limit easily.  The OP's desire to use a third card in addition to the dual SASLP cards may be forward-thinking, but as of today it is pointless (at least from unRAID's perspective).  While it would be nice if LimeTech had the time and resources to get every bit of potentially applicable hardware working, I don't see it as a very high priority.  As long as there is something that works (in this case the dual SASLP card setup), I don't think it is too much for LimeTech to ask their customers to use only that hardware for the time being.

 

That said, LimeTech does have a good track record of working with customers to get certain hardware working and the appropriate drivers integrated with future unRAID releases.

The OP's desire to use a third card in addition to the dual SASLP cards may be forward-thinking, but as of today it is pointless (at least from unRAID's perspective).

 

My desire is not a third card, my desire is x8 lane cards as the first 2 cards. I cant fully agree with you because why are unRAID user forced into buy 1 particular card from one vendor? to achieve 20+ drives? Maybe this should be publish on the main page if you wish to go over X amount of drives you have to buy SASLP cards. And then they should buy up all the SASLP cards and sell them with a PRO license, with 2 options 1) Pro License with 1 SASLP card 2) Pro License with 2 SASLP cards with this all being at discounted prices, lower than what anyone of us can buy them for personally. My understand is when you state SASLP you are referring to the SuperMicro cards, if that is not the case, I am sorry.

 

While it would be nice if LimeTech had the time and resources to get every bit of potentially applicable hardware working, I don't see it as a very high priority.

 

Completely disagree, unRAID is not a free product as others, LIME gets paid to keep moving forward.

 

As long as there is something that works (in this case the dual SASLP card setup), I don't think it is too much for LimeTech to ask their customers to use only that hardware for the time being.  I dont see where on the main page LIME ASK's their customers to use SASLP cards?

 

I dont think LIME would agree with your statements here. If any of the big companys stated what you have stated here they would be eaten alive. The statement for the time being, also troubles me, how long has the SASLP cards already been supported for? I believe quite some time so it time to move to support more.

 

Quick question, is LIME doing this for fun and a few dollars on the side or growing???

I agree with madburg on every point he makes.  Limetech IS currently interested in making the BR10i work.  Evidence is in the latest beta where temps are displaying for that card now.  There are actually folks who NEED to see that card supported because the AOC-SASLP-MV8 is NOT supported in ESXi and that is a roadblock to using unRAID on ESXi based servers. 

 

Limetech has made it clear that he wants to provide support for larger hard drives and additional hardware.  Improvements in the code/test/release cycle are extremely encouraging.  We have an historical precedence for supporting more than one affordable 8 port card.  Without warning the 8 port adapter Limetech was using in his own builds was suddenly discontinued leaving the whole community in the lurch.  I believe that lesson was deeply felt by those who have been involved with unRAID that long.  Let's not return to that again.

Just to be clear, I am not trying to pound anyone, Rajahal you are a great contributer and as Queeg pointed out, LIME is trying. But I am a bit fustrated as I dont see a bit of clarity from LIME. The development site has not been update for a while, which I found great to know what they are working on, how far into it. What ideas/requests were taken into consideration, etc...

 

I personally dislike the SASLP cards due to the price point for what you get a x4 lane card with a Marvell chipset. Great x8 lane cards can be bought for less with better supported (Wintel/Linux/ESX)/performance chipsets IR/IT cabablilities. but LIME clearly has to jump on board which goes without saying comes back to them at least 2 fold.

 

My X8SI6-F-O is shipping today and I REALLY hope that onboard SAS 2008 controller works with unRAID. The Br10i is also going in there and unfortunilty for now will not spindown automatically (via spin groups).

 

Rajahal you build many 20+ drive unRAID systems with X8SIL boards, would you not agree just buying a X8SI6 which give a better price point because of the onboard SAS/SATA controller is already included (versus trying to by a new card with a SAS 2008 chipset) and has better specs than a SASLP card. And still have a PCIe 2.0 slot for another x8 lane SAS card. Or add a SAS Expander (like HP, support 6Gbps operation) to one of the onbaord SAS2008 ports. Opens many doors, would you not agree?

Surely you would struggle to saturate x4 PCIe?  Where is the data going?  6Gbps SAS/SATA is a red herring if we're talking mechanical drives since they don't even saturate 1.5Gbps yet in most cases.

madburg and Queeg, your points are well taken.  madburg, you are right that the X8SI6 board would allow for cheaper 20+ drive servers, and with better specs, and no one would complain about that.  At the current time I don't believe (and correct me if I'm wrong) that the SATAIII speeds will help anybody today as no hard drives on the market can use all that bandwidth (hard drives still don't even pass the SATAI specs!).  If you were building a server based solely on SSDs then it would certainly matter.  Even still, I'm all for using hardware with the latest specs so that it will be ready for HDDs that can surpass SATAI speeds once they are ever released.

 

I do hope that the X8SI6 works out for you.  If it does, I'll probably buy one too and start doing some testing.  Who knows, maybe it will become my next 20+ drive recommended board

 

Queeg, I had forgotten about the conflict between the BR10i and ESXi, that is certainly a valid point as to why LimeTech needs to be (and is) working on offering full support for that card.

 

The point I still stand behind is that I don't see much of an incentive for LimeTech to support the latest and greatest hardware just because it is new and flashy.  Many of the features we are seeing roll out today are basically useless in the unRAID world, but can certainly interrupt unRAID operation.  I'll cite the past example of HPA and older Gigabyte boards.  The HPA 'feature' effectively broke many unRAID arrays.  LimeTech basically did nothing about it, and it its stead the community stepped in and advised users as to how to recover from HPA woes and to avoid certain (most) Gigabyte boards in the future.  Could this have been handled better?  Probably.  However, the fact remains that the HPA fiasco was ultimately the fault of the motherboard manufacturer, not of LimeTech.  I for one would rather see LimeTech focusing on improving the unRAID product, adding features, etc. instead of constantly chasing the tail of ever-evolving hardware.  That's what people like Queeg and I are for

Limetech isn't chasing the newest flashy by adding support for additional 8 port cards.  He's adding value to his product.  Value that other users have asked for.  Some of those discussions lead to existing solutions, some lead to new requests.  All lead to greater community knowledge.  I encourage everyone to push the boundries of the forum (and unRAID by extension) into new territory that expands the market.  

 

unRAID and Limetech aren't the possession of any particular segment of the user community.  For every member that posts there are dozens or hundreds of interested lurkers. 

I do hope that the X8SI6 works out for you.  If it does, I'll probably buy one too and start doing some testing.  Who knows, maybe it will become my next 20+ drive recommended board

 

I am SO crossing my fingers as I drop money on a X3470 (with hopes of ESXi and virtualizing unRAID it's basically an i7) and ECC Registered memory. I will start out will just unraid, and then move to ESXi and virtualizing unRAID. I dont know enough about unRAID yet so need to see how happy, stable unRAID is with the hardware first (what works what doesnt) before moving to a virtualized unRaid setup.

 

We will leave the rest alone  , we all support unRAID and we just need a bit more love at the moment from LIME  

 

Hope to work this board with you guys soon.

Limetech isn't chasing the newest flashy by adding support for additional 8 port cards.  He's adding value to his product.  Value that other users have asked for.  Some of those discussions lead to existing solutions, some lead to new requests.  All lead to greater community knowledge.  I encourage everyone to push the boundries of the forum (and unRAID by extension) into new territory that expands the market.  

 

unRAID and Limetech aren't the possession of any particular segment of the user community.  For every member that posts there are dozens or hundreds of interested lurkers.  

 

Agreed, without guys like JoeL and many others, we would not have alot. And we dont pay them.... so requesting that more than 1 SAS controller be supported fully is definitely not to much to ask for, when paying  

 

I dont believe LSI chipsets are asking for to much. No one here is buying $1000+ controllers that would also have to be supported.