New Server!

Just crazy. A while back when I had acquired the Dell XPS 710 I was using (that died) there were also a few other machines that fell into my hands. Most of them where old desktop garbage, under 1ghz 256mb ram stuff, eck lol. But there was also a server. A Dell PowerEdge 840. The machine was stripped of all drives and memory but it did have a SCSI card, which I have no use for. I had to borrow some ECC unbuffered memory from a friend to try and get it to work, which I couldn’t. I was able to get it to boot but nothing I did would make the damn thing recognize a hard drive. Even adding a PCI-E SATA card, nothing. I could not bring myself to throw it away however, I have a hard time letting go I guess. Which is a good thing!

After the server died last month I have been trying to figure out where I am getting the cash for a new one or “what else could I do”. I was going to go down the PCDuino road when I decided to try the PowerEdge one more time. I was pretty sure I did not have any ECC ram, but sure as shit when I looked in my Bag-o-Ram there was two 512mb sticks sitting there. WTF, I do not remember those. So I tossed them in and the machine booted up. Quick look through BIOS, yup everything looks OK. Power down, added a hard drive, powered up. #$%&@#$$%&!!!1 It fucking worked. I did have to enable all the hard drive options in BIOS but I would have sworn I would have tried that the last time too. Fuck it, don’t look a gift horse in the mouth.

Moved all the drives, minus two smaller IDE drives (not enough ports/room), and ClearOS booted. Albiet into emergency mode or whatever but hey it booted. I had to issue a command and no I don’t remember what it was, but it was printed out on the screen. “Hey idiot, do this.” And then a reboot once or twice and ta-da Bob’s your uncle. I did have to do a little finaggling with the ethernet card, but since I moved the PCI-E card too it wasn’t that bad. At first ClearOS gave me all new card names (like eth0, eth1 instead of enp3s0. Which I always thought was an odd name). but it renamed them all to the older scheme after a reboot. It basically only took about 20 minutes after I moved all the hardware over to get ClearOS up and running. I am very pleased – for once.

PCDuino power

So yeah like the title says, the PCDuino seems to crave power when running a 2.5″ SATA hard drive. When using a SATA drive you need a 2.5AMP power supply, MINIMUM, or you will experience issues with the hard drive shutting down and disconnecting. With the PCDuino by itself a 2amp and 2.4amp supply seems to be ok. I was using an Apple iPad charger rated at 2.4amps, I use it on the Pies all the time with no issues, although I have read that others have some problems with them shutting down after a while (can’t find the link right now). I moved from the iPad charger to my Orico USB power supply (rated at 2.4amps each plug). It worked here ok but after 20 minutes the same issue appeared. So I pulled out the wall wart power supply that came with my CanaKit. Its a 2.5amp power supply that shipped with my RPi2. This power supply has been working flawlessly.

I tried two different 2.5″ drives with all of the power supplies I mentioned. One drive says 0.55amps the other 0.45amps both at 5v of course. They would both power up but after about 10-15 minutes the drives would start clicking, like they were losing power and cycling over again. Access to the drives was also lost while this happened. I tried three power supplies that were all rated 2.4-2.5amps. It seems they are not all created equal…

I need to get my hands on a USB current meter.

Side note: I purchased a BananaPi SATA hard drive cable kit off of Amazon for like $3. It will work – but you must SWITCH THE POWER AND GROUND CABLES! The BananaPi swaps the 5v and GND on the power cable. Just pull the pins out and switch them and it will work just fine. Or get a cable for the PCDuino or I head the Cubbieboard cable works too.

Server Death

I lost the server. The motherboard decided to depart from this world while I was out of town the last few days (this post was written a good week or so after the actual event). Good news is that I did not lose any data or hard drives, just the server. It’s probably going to take me a minute to acquire a new server, I can’t just get a new motherboard. I was using an old Dell XPS 710 that apparently uses special BTX connectors so no replacing anything. It was old anyways. I do see two swollen capacitors so I may try to bring her back, we shall see.

UPDATE: I did swap out the two swollen caps that I saw but it changed nothing, the board still did not boot.

I need to backup my DB a little more often it seems, I was only able to go back to October. “Lost” two months of data. I say “lost” because it’s not really gone it just needs to be retrieved from the (working) drives.

I happen to have a few Raspberry Pies and a PCDuino Nano sitting here on the desk so I commissioned the RPiv2 to be the new temporary server until I can get my hands on a new one. The Cisco E1200 that I had as an emergency backup router is in place so the house has network access again and the Pi should be able to handle Apache and OpenVPN for a while. Sadly I have lost all my custom firewalls and other roll-your-own-server/router goodies. Back to being a “network civilian” again, shitty.

You wouldn’t believe the story behind this server anyways. When I got the machine it had been stored in a barn for a few years. It had about 1″ of crud inside it that I had to clean out and it had apparently been in contact with water as there was evidence of such. I removed the sound card as it looked damaged by water and other crap. But to my surprise after a cleaning it fired up. After a bit of testing it was decided that it ran fine so I moved it into production – and gave away my old (loud as fsck!) AMD Phenom… doh! Turns out three months later that was a bad idea. Two days ago I found out this Dell had once been sitting on the bottom of a pool. Whoa wtf!?