Since 3ware controllers are not supported by Dell on Dell servers, and as this is a very low-end server, there could be many different things at fault - the controller may be requesting something of the BIOS that the BIOS is not capable of providing or doesn't know how to deal with, the system may not have the resources to allocate to that specific card's needs, etc.  I'm not sure what else you might be able to try - remove any expansion cards other than the controller to see if that makes any difference.  

The 3ware 9650SE also has not been tested by LSI, which doesn't necessarily mean it won't work, but that it "may" not:

www.3ware.com/.../Motherboard_Compatibility_list_9650SE_900-0026-02RevK.pdf

You may also look into the possibility of a more recent/possibly compatible firmware update for the controller.

Nobody?

I would expect it to support PCI cards capable of operating with the CPU and chipset.  As mentioned in my first post I updated both the controller and server firmware.

The card works in a HP Proliant MicroServer which is very 'low end' and cost a fraction of the T110II.

My past experience with those cards has been that the Physical Memory Manager cant locate a contiguous block of memory to load the manger into. On custom built boxes I was able to move the memory blocks around, but I dont think thats an option on the T110ii. as a last effort, try disabling everything possible in the bios to see if that will free up a block. Good luck.

The N36L - that was around 8 months ago. It cost me £130 with 8GB RAM (including £100 cash back).

The T110 II cost £506 - it would have been less but the pointless 250GB SATA disk was mandatory.

So it sounds like Dell knobble the BIOS to stop us using any RAID cards other than the S100/300 and H200 on a low end server?  I was planning to replace the 3ware card with an Adapted 5805 but it seems likely that won't work either..  Beginning to wish I'd opted for the HP ML110 instead - I doubt that suffers from this issue.

A fraction of $399?  I guess $349 is technically a fraction (7/8), but just the difference one would imagine when you say "for a fraction".  If you are talking about the N40L, then we are also talking about the difference between an 18-20 month old AMD chipset, versus a 5-8 month old new line of Intel chipset (C200-series).

3ware seems to test motherboards with specific CPU's, chipsets, and BIOS versions and not just specific CPU's and chipsets.  The fact that they stopped certifying/testing motherboards 2 years ago may mean something as well.  Perhaps boards with newer and newer chipsets do not properly support that particular card.

Good luck.

Many vendors have a nasty habit of spending $ to remove features from their hardware thereby crippling a computer that people have paid good money to own. The fact that these vendors do not document the crippling, coupled with poor knowledge by their support staff, makes it rather difficult for an owners of the products to know if they are facing a genuine fault (which should be fixed) or something purposeful by the vendor (which obviously they will not  fix). It's sad that so many vendors behave in the same way, Dell are not alone here.

The best approach to take when confronted by such issues is to contact the vendor and ask for a technical explanation of the failure. If no technical explanation is forthcoming or the statement is that a specific card has not been certified for this machine, then this can only be interpreted that the vendor is being incompetent at best, disingenuous or deceptive at worst with respect to the systems capabilities.

Being that PCIe slots are built to a industry standard, this brings an expectation that all PCIe cards built to this same standard will work. So one would rightly feel justified that the product was misrepresented if indeed it didn't function as the industry standard dictates and any right minded person would expect. In such a situation a full refund would be due in many jurisdictions.

In this case, the only way to not misrepresent a computer would be for the vendor to make an explicit statement that the PCIe slots are not to industry standard and only vendor certified PCIe cards will work but people wouldn't buy such restricted computers hence the lack of such statements.

This type of behavior needs to be jumped on by all for it to end. The one thing all vendors understand is the bottom line, so, if in these situations we demand and get a refund for these misrepresented crippled systems, then go to their competitor for something that has not been crippled, the situation can and will get better. If we accept such products and don't return them, the vendors are only emboldened (an unjustifiably enriched).

So, as i see it, in your case it's either a fault and the BIOS needs to be updated to correct the fault OR the system was designed in a crippled way and you have been stooged and should return it for a full refund. Then you can buy the HP ML110 if it better fits your needs.

PS. I have bought a few Dells in my time but the last Dell issue i had was rather poorly handled by sales/support. This week i bought my first HP, a Z210 which i just received, and despite the 2 issues i had in the first day, HP support fell over backwards to resolve them both within 24 hours. So far i am happy with this machine.

skylarking ... we have had some friendly discussions about this before - and I don't want to turn this into another one, but I wanted to point out one thing:

"vendors do not document the [base hardware feature/functionality modifications]"

"coupled with poor knowledge by their support staff"

Sorry for the edit, but in a small percentage of the time is the modification a marketing ploy - it is usually done for stability (an OEM's primary goal, or they don't make any money).

The sentences should be reversed:  "poor knowledge by their support staff" BECAUSE "vendors do not document the [base hardware feature/functionality modifications]".

The reason that the support staff SEEMS to have a poor knowledge of these modifications is BECAUSE they are NOT documented - internally (beyond engineering) or externally.  It isn't a big conspiracy that they remove functionality and it is a big joke within the company that they talk and laugh about amongst themselves; it is that the support staff is never told of such modifications, and in many cases (as they are not the primary USERS of the equipment), they have no idea what the base functionality even is/should be.  Support staff is given the specs, information, and knowledge to support the features and functionality that the machines are designed for by the engineering teams.  Engineering is the only group that would be able to shed light on reasons why some base functionality never makes it into a final OEM product - be it a marketing strategy or for technical reasons.

to add to what Flash says, I can promise you that Dell would not spend money to defeature a system. Any company wants to make a profit and we are no different, so spendning money to make a system less robust isnt done here. I have been a Dell for many, many years and seen alot of things, but this wouldnt be one of them. The part that seems to get lost within the company (Dell) is that if a box should be capable of something (specification wise) Dell will provide best effort to make something work and at the very least we should tell you why its not going to work. What we usually ask for is your hardware to duplicate the issue with. We dont usually go buy the hardware since this could get very pricey if we had to buy hardware or software everytime a customer claims something doesnt work. This is where we usually lose the customer interest since they dont want to send us the hardware. Not supported means we didnt test it and we dont have the knowledge to tshoot it outside what we know the system is capable of, but if its a standard device, we should be providing "best effort" to understand why its not working. I suggest to Justin that he open a case with support if he is willing to send the card in and they should escalate the issue up the food chain to Engineering to determine the issue. I see it as an opportunity, if this card doesnt work, what other cards may not be working and we could be losing customers because of it, which is a negative.

To update on this - ESXi was still able to see the card and I configured it via the tw_cli command line program.  It's not as convenient as the BIOS utility but it will do.

With regards to the BIOS memory issue.  How can I submit this issue to Dell so that if the issue could be addressed in a future BIOS update it is given consideration?  The only way I seem to be able to send an email is via a warranty request which doesn't really apply.

As long as you are willing to send a card in, you will probably need to call into tech support and tell them you want engineering to root cause why the card doesnt function properly and that you are willing to send the card in. They should then open a case to the Product Support guys and it should go up the food chain. Dont let them tell you otherwise, ask to speak to a manager if you have to and you can quote me on what I have written in my previous post.

theflash, changing my use of 'crippling' to 'base hardware feature/functionality modifications' only makes for a politically correct sentence but that doesn't change nor address the issue itself. That is, why is it that system limitations are not more accurately communicated by Dell in their documentation of the system?

It's not correct for Dell to exclude everything as unsupported by default unless it's an option in the ordering process. Expectations are such that devices using established standard interfaces should work. If they don't, there must be strong technical reasons why and such limitations must be documented. Not documenting any limitations is just misleading the customers!

One should rightly expect that if some document states that the PCIe 2.0 slots are provided, then all cards meeting that spec would and should work. If, as part of that spec, some Option ROM size is defined, then if Dell chooses to reduce the available ROM size during design, bringing it outside spec, then the only correct thing to do is highlight such a limitation in their documentation. But Dell do not document as they should. I have fallen prey to such poor documentation (of much simpler is standards) in the past (and the unexpected issues that resulted). So to me it's not relevant why Dell doesn't have its house in order, just that it needs to lift its game in this regard. If they need to fix their internal documentation and support training so be it. But in all instances their customer documentation needs to be improved.

And everything that is done by a vendor is not always about stability. If it was, raid cards and enterprise drives would have had their protocols standardized by all vendors years ago and raid cards and drives would self optimize and vendor certified drives would not exist today. (Atleast on the perc6/i front, Dell did the correct thing and allowed uncertified drives to be useable on these raid cards). It's all about the $ which is ok in some ways but vendors shouldn't hide the truth when and if they chose to limit or 'cripple' a design. Maybe on this point we can agree to disagree.

Remember IBM made the choice back when it designed the PC that it be open. Now all vendors are trying to put the genie back in the bottle so to speak and are pushing to lock down the architecture. If we're all not careful, we will soon have a chain of 'trust' from bios all the way to the apps that will tie us to specific vendors at out detriment while viruses will still exist. otherwise the PC will wither and die and we will all be worse off.

Sky, it sounds like you are suggesting that Dell is omitting information on purpose and thats not the case. We try to document problems in the User Guide in the FAQ, but due to scheduling, there are times when we have to put things into readme's.  We do document scenarios that customers could run into, and most of these are published for our tech support consumption, and some of the articles are made public as well. We are working on the process to make more of the articles publicly available, but this will take time and a change of views within the company. In the case of this card not working properly, we will make the attempt to understand why its not working and go from there :emotion-2:

The only ones that know what information is being excluded and why is Dell. Obviously some information would be defined as proprietary and may not be made available to support staff and this can be understandable in some instances. But to offer best in class support, support needs to have access to information - accurate information (and sometimes detailed technical information).

My gripe is that even basic information as contained in the Hardware Owners Manual is in some instances wrong and little if anything is done to correct the document (even when the errors are highlighted to both sales & support). To have late updates placed in ReadMe files is OK as a stop gap measure but the base documents should be revised at some later time (it's all PDF's these days so it's not like we actually have to print updated booklets and throw away the old ones like days gone past). Unfortunately these updates never seem to happen which is not good practice. Also, some information that should be supplied in the HOM is lacking. For example my first HP purchase, a HP Z210, has all connectors and pins identified in the manual. Try getting such information from the Dell HOM or other available documents! When you question Dell on this, it falls on deaf ears.

Dell needs a wake-up call and in part this is why i am critical in many postings. And it's great to hear that Dell is trying to improve by making more of the (internal) articles publicly available , but the base documentation must be accurate and contain more detail, that should be the starting point.

If Dell management read some of these posts more regularly, they could see a need to improve and their view would change more quickly.

As for the OP's problem, he seemed to run into a wall so the support process and options must be made more visible when issues are logged, otherwise Dell  frustrates their customers and creates bad will rather than good will. I'm happy to hear that OP's issue is going to be looked at in more detail and hope he posts how it was resolved (firmware update, etc).

Agree with skylarking on most points, the technical documentation is poor, the technical support is even poorer.  No way to get in touch with a technical support contact other than this forum or a warranty based request which isn't really relevant.

I'm quite disgusted with the initial response being that the issue is probably because this is a "very low-end server".  Do you really expect me to spend $$$$ on a higher end server for home use particularly as they come with old fashioned CPU options at twice the price?  This is very obviously a fault with the BIOS memory allocation which I'm sure can easily be fixed in the next BIOS release but with no sensible means of reporting it or submitting diagnostic information I very much doubt that it will be.  Do you really expect me to buy a second RAID controller (which costs more than the server) and send it to Dell so that you can 'possibly look in to' resolving the issue?  I don't think so.

I won't rant on about this, in the end the card worked even if I can't configure it via the BIOS interface which makes life easier in the event of an OS failure etc (no assistance from Dell was required, I just needed to install ESXi and use the 3ware cli tool to configure the card rather than the BIOS utility).

I do think it's short sighted of Dell to adopt the "well it's a cheap server what do you expect" type of approach to supporting these things as they haven't taken in to account that as a senior manager for a service provider I have responsibility for purchasing considerable quantities of hardware for both our own use and that of our clients.  Something like this doesn't really leave me with a positive impression of Dell overall when I compare it to the experiences I have had with HP.