Hi,

I have the following case:

There is + - 30 days I caught a server in The Planet, a DELL PowerEdge SC1425, Dual Xeon 3.2 2 GB RAM, 2 HDD 230GB SATA, to replace an equivalent server that was in other datacenter with backbone problems.

The server is presenting both I/O wait and load average very high, ending up to passing of 90% in determined moments. The simple fact of typing TOP (after have stopped all the services, cpsrvd even), already elevates I/O for more than 15%, coming back to 1% less then 1 second. With all the services running, I/O average stays around 30% to 70% and load ended up to passing of 100, and the usage of RAM is less than 30% and usage of swap, 0.1%. In WHM, in Server Information, the details does not appear of HD, just, sda: Write Protect off is. There is no errors message in the logs generated during boot.

I initially thought it needed to enable something in the kernel (since stock comes optimized for compatibility and not performance) and I compiled a new kernel, incapacitating the support to IDE and enabling the support to LOW LEVEL SCSI - SATA. But nothing.

Look hdparm:

root@server [/]# hdparm -c /dev/sda

/dev/sda:
IO_support = 0 (default 16-bit)


when I try setup to 32:

root@server [/]# hdparm -c1 /dev/sda

/dev/sda:
setting 32-bit IO_support flag to 1
HDIO_SET_32BIT failed: Invalid argument
IO_support = 0 (default 16-bit)


About the possibilities to of being some services setup, already discard, because in the other datacenter was working normal, and the setup is the same.

Could the solution be in the activation of some BIOS's Argument ?

Thanks.