Server wont boot. :'(

[Gaz]

What a great start to the morning. Our one and only DC decided to stop functioning without warning, Its a Dell T310 and the diagnostic lights on the front seem to rest on #2 and there is no display on the monitor.

That error code leads me to believe its related to an expansion card and the only expansion card is our storage adapter, there is also a little green LED on the storage adapter that is flashing. The Fan is also spinning at full RPM to indicate a fault.

 

I've logged a case with support but I didn't have much to give them as there isn't much to go on, no BIOS beeps etc. So if anyone has any ideas that could help I'd appreciate it.

Cheers.

[Arreks]

You are correct in that #2 only refers to a possible expansion card fault. Have you taken the troubleshooting steps in the owners manual? that's all the support will tell you to do.

[Gaz]

Support are coming down to site to take a look they haven't advised me to do anything yet. I was wondering if anyone had experienced that error before. Because I'm not getting a display leads me to believe its either the mother board or the cpu. If the expansion card had died I'd it should still POST.

[Kenny_G]

We had a dell T310 which wouldn't boot, turned out to be keyboard. We plugged a different keyboard in and it worked. I know it sounds strange but it's worth a try we also had the diagnostic #2 light on.

[Sagima]

Can you boot a dell diagnostics USB and see what it tells you?

[Gaz]

Sorry for the late reply. To say its been busy today would be an understatement. :/

@Kenny_G I did try that but to no avail.

@Sagima it wasn't getting a display at all so no chance of booting to anything.

 

Support has been down and think its a duff motherboard which confirms my suspicions. Although there is still the potential for a duff CPU.

[win]

Try pulling out some RAM sticks (long shot but it might work).

 

You don't want to hear this, but DC's don't require much resources so having a second domain controller on the network (even a hyper v one) wouldn't be too much trouble, and could be a life saver. They're easy to setup because it just gets the replication details from the existing DC.

 

Don't suppose you have a backup of this DC somewhere?

[Gaz]

Try pulling out some RAM sticks (long shot but it might work).

 

You don't want to hear this, but DC's don't require much resources so having a second domain controller on the network (even a hyper v one) wouldn't be too much trouble, and could be a life saver. They're easy to setup because it just gets the replication details from the existing DC.

 

Don't suppose you have a backup of this DC somewhere?

 

Yeah lessons are learned all the time, I will be looking into setting up a secondary DC. All but 1 stick of RAM was removed during the testing, with the same result. Error code #2 with no display and fan at full RPM.

 

The Data for the DC should be fine if its just the motherboard. I do have a backup of the DC but hopefully I won't need to use it. If it were near the summer holidays I'd be tempted to upgrade it to 2012.

 

Just waiting on the new motherboard to arrive.

[Gaz]

Server is backup and running. The engineer while waiting for a replacement motherboard decided to reseat the CPU there was a little bit of dust around the ZIF socket, once he removed that and reseated the CPU the server booted.

No idea how that could have stopped it from booting.

 

In the event log in the before the crash are lots of Event ID 1004 ipmidrv messages, a quick google reveals no solid answers to a solution but there maybe a firmware update for the iDRAC

[zag]

I've always used very simple servers for domain controllers for exactly the reason to prevent stuff like this.

 

All the components are instantly replaceable in situations like these.

[plexer]

Quite a good idea use a basic commodity machine with multiples or virtual additional ones for this case and use the high end hardware for other roles and data protection.

 

Ben

[Gaz]

The budget for the server at the time wasn't great and the best way was an "all in one" approach to the server. I'll definitely be looking for another server to act in a failover role.

@zag when you say "instantly" do you mean you can source the parts yourself i.e. from a retailer or do you have them in stock? @plexer do you have an example of such a spec?

 

Cheers.

[3s-gtech]

Glad to see you're sorted. I got a bit DC paranoid back when I took over my current role - realised there was only one DC for each of our two domains (root and child). I used very old Pentium 3 powered stuff until budget permitted better - was still Enterprise grade and very reliable as it turned out. Replaced with an HP Microserver and VMs in the end, but total cost was very small - only the OS licenses that really had any impact.

[Gaz]

This looks like a good deal, FUJITSU TX131 Intel Xeon E3-1226v3 1 X 8GB 2 X 500GB NOT HOTPLUG 3.5 INCH DVDRW | Servers Direct although I don't know if Fujitsu servers are any good.

[plexer]

Any machine you can install a server os on really or just cheap commodity servers like the one you've listed or perhaps the HP ML110 range as they're solid little things.

 

Ben

[zag]

We run our DC's on Dell R320's. Cost was about £800 each from what I remember.

 

All our servers are the same so all are interchangeable in case of emergency.

[marsdenprimary]

Unplug everything, hold the power button down for 30 seconds to drain all power, plug everything back in and try that. Sounds simple but it could work!