Reminded why I hate PC

I used to use used/obsolete Sun SPARC systems as my home server.  The last was an Ultra 60 retired about 3 years ago.  I always bought two so I’d have a spare in case of hardware failures.  About 3 years ago I caved in and switched to x86.  I built two fairly hefty servers that I planned to have for a long time.

While specifics don’t really matter as far as this post is concerned, here they are for those with interest:  SuperMicro H8DM3-2 motherboard, SC825TQ chassis (2U), and SIMLP-B+ LOM (Lights Out Management) card.  Single quad-core AMD Opteron Shanghai 2376 (2.3Ghz) processor and 8GB of Kingston memory.  Dual ST3250310NS drives (RAID1) for O/S and dual ST31000340NS drives (RAID1) for user files.  Currently running the latest Solaris 10 and using ZFS for storage.

A few weeks ago I decided to start playing with Solaris 11 on the spare with the intention of switching to it in the future.  (I had been playing with Solaris 11 Express on VirtualBox on my Mac.)  The spare system had been on and off over the few years I’ve had it – never had a problem with it.

I turn it on only to be greeted with a black screen, no beeps, and all fans.  The LOM claims the CPU core voltage is zero as is the battery.  So I dig up a new battery and drop it in.  Now I get a BIOS screen up to the point where it says the CMOS checksum is bad – press F1/F2 to SETUP or Load Defaults.  Except neither F1 or F2 do anything.  Not from the real keyboard or via the LOM.  Neither does DEL or F4 (LOM mapping to DEL).  So I contact SuperMicro support and get a really helpful response – USB controller must be messed up, try PS/2.  I don’t have a PS/2 keyboard!  I try another USB keyboard – same nothingness.  I find a USB to PS/2 adapter; still nothing.

Finally, out of desperation, I yank out the LOM card and power up.  Surprise!  Everything is working again.  I put the LOM card back in – and it’s still working!

If this had been my primary system and had I not had a spare, I would have been REALLY upset over the days or weeks it would have taken to get this fixed.  Not to mention my really unhappy “users” (rest of my family).

My intent isn’t to pick on SuperMicro.  My experience has been that they aren’t any worse (or better) than any other x86 piece of junk I’ve come across in my career.  I have yet to experience a brand or model that is what I’d call good.

Actually, I’m wrong about that.  I have 4 other x86 systems in my house that ARE what I’d call good.  They all have the Apple brand on them.  So it isn’t correct to blame x86 – Intel or AMD are not at fault.  It’s the fault of all the vendors other than Apple that continue to build mediocre motherboards that still use that piece of crap called BIOS.

BIOS needs to die – and it won’t be soon enough.