CentOS 7.1 rendering hardware inoperable?

[mikedn]

I have encountered a bizarre problem which I have never encountered in my numerous years of Linux use, since the days of Redhat 6.

I have two Dell R410 servers. They were running CentOS 5. I performed a fresh install of CentOS 7.1 on both simultaneously (have installed CentOS 7.1 on other systems, just not this hardware, wasn't expecting any problems).

Install ran fine. Rebooted. Accepted license. Did a YUM upgrade. Upgrade successful. Changed default target to multiuser (didnt need to gui these are principally just file servers). init 6 to reboot.

Only they didn't reboot.

The systems shut down, and appeared to reset. Then nothing. Just a black screen, with an amber flashing status light. No "Configuring memory..." like I normally see. After a few minutes, the internal fans slowly spin up to a higher speed. But no video, no console interaction. nothing. Completely DOA.

This has happened on both R410 systems. They are now unbootable. I cannot get any activity from them whatsoever.

An upgrade performed on an R610 system at the same time as the R410's completed successfully with no problems at all and is running fine.

Has anyone ever seem a similar type of problem? I'm getting ready to call Dell, but both systems are out of warranty, so I do not expect much at this point.

[mikedn]

Just spent over an hour on the phone with enterprise support. They are clueless as to what may have happened to both servers. My only option is to reinstate the warranty ($$$) and swap out the motherboard. $962.47 per server.

[gerald_clark]

If you can't enter the BIOS, it is not a CentOS issue.
If you can, boot DVD in rescue mode and check your drives.
Perhaps your RAID controller is not supported.

[mikedn]

Yes. Clearly this is a hardware issue. Both motherboards are dead. The raid controller is not an issue. As part of dell troubleshooting, we removed all components other than the bare minimum (one CPU and 1 RAM card) to try to get the system to boot. No go. Enterprise support says the only option is to swap the motherboard as they have diagnosed the motherboard DOA.

It seems hardly a coincidence, though, that both motherboards simultaneously STB immediately upon rebooting after a CentOS 7.1 upgrade, though, don't you think? Sure, if one system crapped out immediately after reboot, I could chalk it up as a faulty motherboard which finally failed.

But two, at the same time, within moments of one another? (these are in server room with a temperature-controlled, line-voltage regulated environment, so this wasn't a random power-spike that just happened to hit the moment the systems rebooted.)

I'm curious if anyone has encountered a similar problem. Especially with Dell servers, particularly the R410s. I am hesitant to do further upgrades because of this.

[mikedn]

From bad to worse.

The R610 I upgraded this morning to 7.1 has now failed. It froze shortly after 3pm. The LCD display on the front panel indicates "VTT regulator failure - Reseat CPU". Reseating the CPU of course did nothing. This system is now, likewise, unresponsive and no longer POSTs like the two R410s.

3 upgrades, 3 hardware failures. This is hardly a coincidence.

[TrevorH]

How is CentOS 7 meant to have caused a voltage regulator failure? Be reasonable! The chances are that the Rx10 machines are just old and were failing anyway.