Tech Blog‎ > ‎

DL380 G6 - Unexpected Reboot

posted Oct 25, 2009, 7:03 PM by Victor Zakharov   [ updated Feb 8, 2010, 5:37 PM ]
I noticed many people on the net are looking for this review, or at least for anything that could prove they are not alone. Well, guys, here's what you wanted to hear: yes, our 9 x G6 DL380 servers kept rebooting unexpectedly for quite a while. Not all at once, but still annoying. It all happened at random until we contacted HP and their tech told us to update firmware.

That's right, you need to update everything you can: iLO, RAID etc. Better use latest version of PSP (we used 8.30), it will probably save some time. The good stuff is that after this update unexpected reboots disappeared. But as you probably have already guessed, there is always a bad side. In this particular case we now have problems with rebooting our servers. What happens quite often is that we see either a black screen (nothing happens) or a red screen of death, and OS never loads. There has been no cure so far, we just keep rebooting and at some point it will boot the OS. Luckily, it's quite stable after that.

Now, what's the red screen of death? Personally, it was the first time for me to discover such error. When it happens you see a CPU register dump on the screen. Yep, all those EAX, EBX, ECX, EDX and all other registers along with their values at the moment error occurred, on the red background. You cannot do much at this point, just keep rebooting and you'll get the OS up and running. Be patient, might need 4-5 tries in some rare cases. All in all, not that bad. But I don't like HP for this.

P.S. I might still have that screenshot with the red screen of death lying somewhere. If I ever find it, I'll post it here. For now, be patient with HP servers if you have G6, and don't rush to throw out your G5's as they seem to be more stable, at least according to our experience.

EDIT - here you go, just as promised:


Beautiful, isn't it?

EDIT2: HP posted a solution to this problem (12/11/2009). You need to upgrade your firmware:
http://h20000.www2.hp.com/bizsupport/TechSupport/Document.jsp?objectID=c01955503
I did not test it, because our servers are running more or less fresh updates and they no longer reboot, if already started.
Comments