Gen8 Hardware failure?

[BAcon]

Hey, I am having some issues with my microserver that perhaps some of you may be able to shed some light on for me.

I'll start with the setup and go from there.

HP Gen 8 microserver
6GB of ECC mem (2GB + 4GB)
1 x 80gb 2.5" in the optical bay
4 x 3TB WD Red drives using the B120i controller

I have owned the server setup for around 2 and a half years running Xpenology, fault free up until a few months ago.
Initially I set the system up with all the drives in JBOD, I wasn't bothered about redundancy so thought I would just go with the simplest setup when I was trying to get to grips with setting up a server for the first time.
After about 2 years fault free service it crashed several times, I was able to repair the data drives and carry on, until the array crashed again. This time I was no longer able to even recognise the drives with the server.

So I took each one out and scanned it with the WD lifeguard program on my PC, 7-8 hours a drive! All were apparently fault free, so I RMA'd them all to Western Digital.

I now have brand new drives, they were all recognised by the server, I set it up with OMV this time, setting the drives up into a Software 0+1 Raid setup, all was well for 3 days then the server crashed again, leaving all the drives now unrecognisable to the server (both in the BIOS and in the OS).

My limited experience with servers and linux isn't helping much, but I have always been fairly competent with the hardware side of things, though it has been a fair few years since I have been "into" computers.

I am thinking that this seems like a hardware failure on my servers motherboard, perhaps the raid controller that looks after the 4 bay slots?

Anyone else had any issues with the gen8?

Cheers in advance

Mike

 

[BAcon]

Ok, a late update, I ran memtest on both sticks of ram together and individually and they were fault free.

To back this up I ordered 2x4gb of memory and tried that as well, all fine and dandy.

However my hard drive bays, 1-4 all seem intermittent still. Yesterday the front light changed to red instead of blue, which I believe points to a power issue. So I'm starting to think my PSU is packing up, replacements aren't cheap to check...I really can't be bothered to tear my main pc apart, to check with my pcs PSU.

Might be time to replace the server, given the sky high prices of the gen8 and spares.

Unless anyone has any suggestions? Cheap ones that is!

 

[BAcon]

Yeh I've been looking at the Gen10, it appears to lose iLo though and has a soldered cpu, I always liked the idea that should my server need some more oomph I could throw a Xeon in there.

What is the ML10 Gen9 like? It appears to have a cheap Intel processor but at least it's upgradable? Has iLo etc. My server sits in an unoccupied office, so as long as it doesn't wake us up in the next room the noise level should be ok.

I'm currently giving an old pc I have by my feet a go at powering some of the drives to see if this helps. In doing so I noticed I have my 2.5" drive installed in the ODD port, but connected via a molex splitter, then to a molex to sata connector. I didnt power the 5th drive via the floppy power connector so all 5 drives are on one cable. I am secretly hoping this cheapo splitter could be the cause of the instability, wishful thinking is more likely, but I can sit and hope everything is ok now I have removed it...

Time will tell...Probably the next 5 minutes..

 

[BAcon]

Ok, I pulled the molex splitter that was powering the 5th ODD drive from the main molex connector for the front bays, ordered and fitted a "floppy to SATA" connector to power the ODD hard drive separately.

It's currently without issue....omv is now running, with 3 drives pooled with unionfs and snapraid using the 4th disc for parity.

It literally went from hard drives disconnecting (presumably powering down as I could hear them powering up repeatedly) every few minutes to total stability. Maybe that 12v rail was overloaded. What doesn't make sense is the previous 18months or more of uptime where I had no issues

Oh well, watch this space!