SMP Problems with 4.8-RELEASE

[Jeff Love]

Elliot Dierksen writes: 

> Hello, 
> 
> I have the following hardware configuration. 
> 
> Tyan 2466N motherboard, 1.03 BIOS, 2 Athlon-MP 1800+ CPUs, 512MB registered
> ECC DRAM, Onboard 3Com NIC (xl type)
> Radeon 8500 AGP video card
> Adaptec 3210S SCSI RAID controller
> Comtrol RocketPort 8 port PCI serial card 
> 
> I have started to experience some odd system lockups in the past couple
> months, and it seems to be getting worse. I particular, this seems to happen
> when the SCSI controller is very busy. I have a file system where I store
> dumps of the other file systems. On a weekly basis, I run these dumps. What
> has started happening is the system will lock up completely during one of the
> dumps. When this happens, the system is totally unresponsive. Even at the
> console, it is dead to the world. However, the status lights on the RAID
> controller seem to indicate that it is happy. The system has to be power
> cycled at that point. I have been unable to generate any debug information to
> help chase this issue. 
> 
> I have also had some issues with mozilla locking up and it could not
> be killed. When I exited from X, the system would lock as described earlier. 
> 
> As a desperate measure to try and get a backup, I built and installed a
> single CPU kernel. That seems to be resolving the problem. However, it would
> be nice to bring the 2nd CPU back on line. I looked through all the BIOS
> options, and I don't see anything specifying the MPS version. I would be
> happy to assist in troubleshooting, but I really don't know which direction
> to go at the moment. I have attached both kernel configuration files and
> dmesg output from both during the boot process. 
> 
> Any suggestions??? 
> 
> Thanx, 
> 
> EBD
> -- 
> Elliot Dierksen          "They that can give up essential liberty to
>                           obtain a little temporary safety deserve
>                           neither liberty nor safety" -- Benjamin Franklin 
> 
> Home) ebd at oau.org  (407) 850-9760

Elliot,
I had a similar problem with a new dual Athlon2400 I have built. The problem 
ended up being a thermal issue. I repositioned some cards and fans to 
provide better air flow in the chasis, and the problem has not reoccured. I 
did not try running only one CPU. I had the same lock-ups you describe. 

I'm not sure this is indeed also your machine's problem, but you may want to 
monitor your thermal sensors. These Athlon chips seem to be hot runners. 

<
< Jeff Love
< Burgh-Com Gaming // www.burghcom.com
< MIG #1646
<