[cobalt-users] Ram upgrade may lead to hardware failure

[Carrie Bartkowiak <ravencarrie@xxxxxxxx>]

*crawls into the forum in fatigues, with war paint smeared on like a clown and looking like she's been beatup by a big group of nasty obsessed basketball geeks...*

Hey all!
Just a heads up. Yesterday my data center performed a partial upgrade to the RAM on one of my servers. There was nothing wrong with the RAM they put in. When the server came back up, it was quite evident that the WWWThreads message boards (of obssessed KU basketball fanatics) that run on the server were actually *slower* than before - and the RAM had been doubled from 128 to 256. Strange, but we didn't really think much of it - a bunch of users complained though and accused us of actually reducing the RAM.  ;)
Today the data center took the RaQ down again and finished doing the upgrade, bumping her all the way up to 512MB. When they booted her back up, it was just straight-out strange... we could ping it, it would open an SSH2 connection but would not accept a password. No access to the web whatsoever, the entire http server was down (as was the GUI).
The data center switched the RAM out again (in case it was bad RAM, which it wasn't), but when she came back up it was the same thing.

One of the data center techs had seen this before and leaped into the box, saving everything in my /home directory before it disappeared. Thank goodness - because there was a 1600-user message board mysql database there that we could not afford to lose and my most recent backup was a week old. The tech was able to dump my /home directory into a new machine before the old one died completely. He was not able to save the system files, including my numerous personal configurations (which I've spent all night re-doing... hence the fatigues and war paint).

According to the VP of Tech Support, this has happened to at least two other RaQ4s very recently - the hard drives just seemingly get the partition information corrupted or something. I am not sure if the other two machines were due to RAM upgrades or not. I'm inclined to think that they were not, as he noted that both of those machines also had the strange "it seems unusually slow" symptoms before going belly-up.

So this is just a heads-up. Backup, backup, backup your data. Somewhere OTHER than your machine. Having backups of your machine *on* your machine will do you no good if the hdd decides to head for greener pastures without you.
Especially if you notice your RaQ seeming to be slower than normal, or if the case is going to be cracked open for any reason. You just never know.

And keep an eye on your RaQ4's. This is the first one that's done this to me, but I'll be certain to notice the next time one of my boxen gets "slow" for no reason and take action ASAFP.
--
CarrieB
"One chance is all you need." --Jesse Owens