Non-booting SP on X4540 and X4240 systems?

We have now seen two separate systems (one X4240 and one X4540) with what looks like the same problem - a SP/ILOM that refuses to boot (nothing is printed on the serial port, the MGMT network interface never comes up and ipmitool on the system just says that the BMC is not responding/not installed) and a console video output that is totally garbled. I guess both those problems are connected somehow.

My _guess_ is that the console video controller has died and since the SP/ILOM want to grab it's output it somehow hangs during boot. Or vice versa - that the SP/ILOM has a corrupted boot image and since it doesn't boot somehow makes the console VGA behave strangely.

I'm curious if this is a common problem that many people have seen or if weare just totally unlucky with this?

Have anyone found a way to fix this problem (sans a complete systems board replacement that is)?

We can get the system to boot by pressing the power button and an already installed OS seems to work (except that the SP fault logs and stuff isn't available).

- Peter

 

The VGA should be connected to the system itself and not the IPMI tool;
it should start with the BIOS.

I'm assuming you've connected it from power? Did you reset the
CMOS? I'm assuming you used the proper serial cable?

When the system is down, it will not power the VGA so it would not
stop the ILOM from starting up.

You might want to look at:

http://docs.oracle.com/cd/E19121-01...cing-doradi2u-motherboard.html#50581562_46131

So clearly its software can get corrupted.

I haven't seen this.

When did this problem start happening?

Casper

 

Haven't tried the CMOS reset yet. Check the CMOS battery though - seems to have a good voltage (3.0V). The serial cable works just fine with another identical X4540 server (or rather I just swapped the CPU board between the servers).

We just got them - two identical used X4540/Thor servers - so I don't really know their history/background.

Yeah - I have been trying to find a way to reflash the SP software but so far all instructions I've found just basically say to give up and have the SP/CPU board replaced if there is no output at all on the serial port at when the power cables are installed... All other instructions talk about how to reflash using a running SP (or the pre-boot environment).

I've found a reference SOCFLASH.EXE and a bootable USB stick - but that is kind of difficult considering the unreadablilty of the output on the VGA monitor... Tested another route via ipmiflash/pciflash from the running Solaris 10 system but those tools just say that the SP type is not supported.

- Peter

 

You have a VGA monitor connected and you when you start the system
you don't see the BIOS?

(It is possible that the X server leaves you with a mess so you should
start with a powered-down system, *powercords removed!!!* and then
see how far you get then when starting the system)

When did this first happened?


Casper

 

I also have the same problem on an X4540. There's only colorful garbage and periodic blanks on console screen, NET MGT port never comes up even on physical level. Unlike original poster, I do see normal output on serial console up until the memory check finishes, then output on serial becomes garbage too.

Everything suggests it's a hardware problem.

 

Hi,

We have the same problem with several Sun X4540 servers. Did you find a trick to bring the SP alive again?

Is there any possibility to fix the issue without changing the motherboard. It is really complicated to find hardware for these servers now.

Jerome

 

Dear friends,

I have a number of Sun x4150 servers at my university with identical symptoms. Did anyone ever find a solution to revive the SP?

Thanks!

Dan

 

Hi All,

I just had the fortune of coming across this issue when having no other option than using a display to remedy the situation. Surprise, the VGA was doing the same garbled output as you have described. I have found that apparently this is common on these servers and hence decided to open it up. Nice thing is that it has a PCIe slot where you can just put in another video card. In my case I used a Quatro K620 (because it fits). This allowed me to connect a HDMI or Display port monitor and have access to anything really

Model: X4140
OS: Solaris 10 x86

Hope this helps someone!

Cheers,
Dani

 

This is great news, Dani. Thanks for the information.
Did you manage to get the SP working again? Could you please detail what you did? I have several servers that will benefit from this!
Thanks again.
Dan