I have some decent HP Proliants server that come with "HP Smart Array P410i Controller" enabled, I am using these boxes to set up a Hadoop cluster and I know, RAID is for sure a no-no for Hadoop since the application itself will take care of data redundancy and extra intelligence provided by RAID won't be helpful and might turn down the performance. I tried to disable the devices at the BIOS and the box cannot even access the disk afterwards. So I am assuming the controller is sitting between disks and mother board, and we have to turn it on and configure it to "level0" or something like that.
I am wondering what should I do to "disable" the RAID functionality so it will fit into the Hadoop environment.
When, I boot the box, I got into the configuration for the RAID, I deleted the existing logical volume and trying to create a new logical drives. And I have a few questions.
- Why as default, only half of the disks on Box1 only will be selected, I am not a datacenter admin, and I want to understand what does the "port 3C/4C", "Box1/2", "Bay 1-8" mean here.
- Does that mean I just need to check all the disks and select RAID 0 and leave the other configurations as default, it will fit?
- How should I configure the partition so each disk will have its own partition? I am using kickstarter server to install OS, here is a ks.cfg from purdue university but I can only see the mount point where the device is missing based on the doc, how does it know which device to mount?
This is indeed a question could be split into several questions but I think it might be helpful to keep them in one question so the idea is consistent and it will be helpful for some people like me to have a solution at one stop.
