No Booting

Antonis.A,

I would suggest taking the server to the Minimum to Post configuration, that is removing everything from the server but the following

• System Board
• Processor with Heatsink
• 1 DIMM in slot 1
• Power Supply
• Control Panel and Cable 

Removing everything won't allow it to start, so taking everything but what is listed is what we need to do to still allow it to start with the least amount of hardware needed.

Is the CPU currently installed the original or the new one? 

Let me know what you see.

How did you clean the dust out?  Brush, vacuum?

Antonis,..

You may have killed your motherboard  with static from the vacuum.

Vacuuming is one of  the worst sources of chip killing static. your better off with a bristle brush (with ground wire), vacuum hose should not be closer than 1 ft from any computer part., and that is considering the hose has a discharge wire in it., and you are fully grounded.

On the tech's  side, if it is a static issue, the CPU is harder to damage when it is mounted, more likely a sensitive component on the motherboard would be damaged.

No best way to dust computers out but I use  canned air most of the time, I make sure I am grounded either by the case or  a wire to ground, if I am grounded the canned air is also grounded, I used the weakest stream of gas I can as static is created with any air movement over plastic (air nozzles, vacuum hoses), less with lower velocity.  Air spigots from factory air supplies are not good, they create a lot of static. High volume air can be used to clean fans, power supplies, computer peripherals as these items are not very sensitive to static.

A bit late to the party but I'll comment anyway :emotion-1:

Cleaning a server, workstation or PC is no simple task but it's not overly complex either. A good clean requires a complete dissasembly so one needs an understanding of how the system components are cabled together and how the system components are logically configured along with the correct dissasembly and assembly methods (with a focus on appropriate electrostatic precautions so as not to damage the components themselves).

Together, the Dell T110 III Hardware Owners Manual and the Technical Guide Book provide much useful information about your server and how to perform certain actions. There are also warnings about what a user can and should not do (without impacting their warranty presumably).

Unfortunately these documents do not clearly indicate if the T110 III PSU is an industry standard ATX power supply, so unless the technician at the Hardware shop you took your system to knows with certainty, plugging in an ATX PSU may not be a good idea even if the plug fits...

It's also unfortunate that these Dell documents do not list a pin out for all connectors on the system so one can check if the standard ATX PSU pinout is indeed compatible!  Further, these Dell documents also do not provide any guidance if and how one should clean or get rid of the dust bunnies that eventyally choke our systems.

But it's safe to say that using a vacuum cleaner can generate massive static electricity depending on the type of cleaner and the air temp and humidity so it's best to avoid them altogether as already mentioned. It's best to fully dissasemble the system (using static wrist straps) then clean out the empty case with a brush followed by damp cloth. Then clean all the individual components (which should be earthed) using a static wrist strap and an earthed conductive brush. In this way the chance of electrostatic damage is minimal. Just be careful that crud doesn't get into, for example the PCIe slots or other plugs or connectors. Finally you can reassemble the individual components back into the clean case again using a static wrist strap. 

The real issue is why did you heavily clean the case using a vacuum? Was it overheating?

I usually just pick out the fluff from around the fans, heatsinks and grills and then wipe the fan blades down using a damp cotton ear bud.QED and safe.

As far as the non booting system, I would read the HOW and TGB to understand how your system is cabled and fix any obvious mistakes if any. Then I'd remove the CPU fan and heatsink and carefully remove the CPU. Then inspect the motherboard CPU pins with a magnifying glass (or maginifying app or camera on your smart phone if you don't have one). Problem pins should be obvious but if all is OK then clean the old thermal compound from the CPU and reseat the clean CPU and apply new thermal paste. Finally, clean the old thermal compound from the bottom of the heatsink and attach the clean heatsink and fan. If you're lucky the system will boot and you would have saved some $ on a new motherboard.

Note with new Dell motherboads, I think there is some manufacturing data that must be added using some proprietary tools. Second hand motherboards may have the wrong data for your system and may result in the OS not staying activateted though I'm not sure. Likely others could clarify if such issues are of concern and need to be handled in some way.

Good luck.

Hi "Skylarking" and thanks a lot for such a detailed response. Obviously you know a lot about hardware :)

To be honest, I already ordered the motherboard so I will not inspect the current motherboard!

There is some additional info that might shed some more light.

Before I cleaned the server, there was always an amber hardware error light on the LED. I assume it was showing up because i replaced the stock fan with another one. Temperatures were just fine however that light was blinking.

I thought that this was the problem and put back the stock fan but still no boot...

It's best not to ignore a system health indicator amber led unless you clearly know why it's flashing and can determine it's safe to ignore. The same phylosophy should be used for diagnostic lights and system messages during boot, that is don't ignore them unless you determine it's safe to do so.

Unfortunately the HOM or Technical Guide do not elaborate on what to do when the system health indicator led is amber, though the HOM does decsribe actions to take when diagnostic lights or system messages occur dutring boot.

Dell uses a case mounted 4 wire fan on the T110 II. Such 4 wire fans provide a tacho signal which the BIOS reads and uses to verify the controled fan speed is as intended. The fan is case mounted and works along with the cooling shroud and CPU heatsink to cool the CPU and memory. If you replace the Dell 4 wire fan with a aftermarket 3 wire fan, then i'd expect an amber system health indicator led and possibly full fan speed and lots of noise to go with the full fan speed. If you replaced the Dell fan with another aftermarket 4 wire fan that has lower air flow (and lower noise) then the components (CPU or memory) could overheat in some instances. So the real question is why did you replace the system fan and what was the real reason for the amber led?

As to the new motherboard, there are system specific settings like UUID, serial number, service tag and likley others. Some of these pieces of info may be used during OEM SLP activation of the OS and if lacking the OS may eventually be seen as priated. Just don't be surprised if this happens.

Oh, and since you have ordered another mobo, i'd check the PSU is indeed working correctly before plugging the PSU connector into the mobo and powering it up. Since I don't know the PSU connector pinouts for the Dell PSU, keeping in mind the Dell PSU may not be to ATX specs, I can't say how to test it but maybe the Hardware shop you took your system to can help here. Last thing anyone wants is to see a new component fried because it wasn't the old one that was faulty but ratehr the component that connects to it...

Good luck...

Thanks for the input! So here's the update.

I got a new mobo and took it to the repair shop to install it. They did and once we pressed the power button, guess what? The server still gave the number 3 LED warning with no boot / video. I asked him to test the PSU once more and he did and he told me that it was OK and that he doesnt really know what the issue is. He told me that a fried CPU is very rare and that I should try to find someone to lend me a CPU before ordering one.

Went to another shop. They did not have a CPU but they tested the PSU once more. Gave a warning that the PSU was not working correctly. So we found the issue. The strange thing though is that they both tested the server with another PSU (not original Dell T110 ii ones) and the server would not start. Question is, should the server start with a PSU that doesnt have the exact specs as the original one?

In any case, I ordered a new server, since it is cheaper than repairing the old one. 

Another question - the new server that I ordered comes with an Intel i3 processor rather than the Xeon, reason being that it is cheaper. It is still the same exact server T110 ii. Could I take the Xeon that I have in my old server and put it in the new one? What if the CPU of the old one though is damaged? Could that harm the new server?

Many thanks for your precious feedback!

Sorry to hear that your new mobo didn't solve your problem.

But I'd probably find another 'shop' to go to for technical support as the current 'shops' you use don't seem to know Dell servers or whether Dell uses a standard PSU's or some Dell proprietary wired psudo ATX standard (that is not compatible with off the shelf standard ATX PSU's and their ATX PSU testers). In any case theses shops should have had test components so they are able to fault diagnose a little better/more completely than what they did.

So, best to find a Dell Server capable 'shop' or someone that is capable of correctly testing the Dell T110 ii PSU and capable of correctly fault finding (without you needing to incrementally buy new components along the way).

As to the i3 2100 series processors, they are 2 core hyperthreaded CPU's processor while the e3-1200 series are generally 4 core hyperthreaded CPU's with a larger cache. The differences between these CPU's really depend on the type of workload but with single threaded workload there wont be a huge difference. Either CPU should support ECC or non-ECC memory.

I'd order the needed/desired CPU with your new system and I'd also get the on site warranty on the complete package. Then use the Dell supporty as needed when you have issues. There's no point in trying to save some $ by using a suspected faulty CPU and possibly compromising the warranty of your new system.

Good luck.

I would have to agree with Skylarking, which he said well.  The onsite support option would be the best route. Also, depending on the system you get there may be no compatibility between the processors for the system, so as Skylarking stated it doesn't save you time or money to experiment on the new system.

So here's an update!

Finally I got a new server. I calculated the costs in replacing the parts on the old one and it made much more sense to simply get a new one, especially since Dell was running a promo on their site.

Now I got a question. In my previous server, I replaced the stock fan with a Noctua NF-A9x14 PWM. It was working find for several months and was running almost silent. I don't think it was the fan that damaged the server since it was running for almost a year with no problems and/or overheating.

Question 1: Could the fan have damaged the PSU? I understand that the PSU has a fan of its own so it wouldn't depend on the main server fan right?

Question 2: Would it be safe to put back my Noctua fan into my new server? As I said, temperatures seem to be the same as with the stock fan, but the reason why I prefer the Noctua is because it runs almost silently due to its design and technology. Would you recommend me any specific steps and/or precaution measures if I put in the Noctua fan?

Many thanks once more!

I'd consider keeping the server stock unless the fan noise if too much.

As to Q1, the system fan would not result in damage to your PSU since the fan is powered by the mobo itself. The PSU has it's own fan and this fan would ensure that the PSU airflow is kept within the designed levels.

Q2 is harder to answer as one should know the designed airflow rates for given fan PWM voltage verses system/component temperature when using the OEM fan and compare this to the same when using the Noctura fan. In that way you would notice if airflow is too low in some instances. But, as you say, if the Noctura fan was keeping the system components cool and you verified this by running fan speed and temperature monitoring software when the system was on and without having issues, then i'd say it should be OK.

Just remember, if you place the Noctura fan in the case and you have heat related issues, you should not expect Dell to fix it for you.So unless the fan noise bothers you that much, leave it standard or test it well if you do the change.

And as always, caviet emptor, carpe dieum and such :emotion-5: