Adding RAID 10 to current setup
 

I've been attempting to get a raid 10 setup going with a current linux install of Centos 5.3. I have set the raid up with the onboard intel driver(jmicron) and it shows as RAID10 when I boot up. Once I get into Linux however and do fdisk -i I get this as the following output.
The 4 750GB drives are the ones set up in the the raid array so I'm assuming the GPT warning is for that. My other guess as to why the drives are shown without partitions would be because I have yet to add any partitions. Would I do that by using gparted and creating a max size partition on each drive? Then of course I would have to add an entry to the fstab, which drive would I put as the source?

The jmicron is supported through the dmraid (Device Mapper) tool. You will need to ensure you have the rpm package dmraid installed.

I would suggest doing.
if it comes back with a package name then you have it installed if not try

dmraid is installed and I just updated it. I'm unsure as to how to use this program though. dmraid -r gave this output:
[homeworld] ~ > dmraid -r
/dev/sda: isw, "isw_dijhgahgdb", GROUP, ok, 1465149165 sectors, data@ 0
/dev/sdb: isw, "isw_dijhgahgdb", GROUP, ok, 1465149165 sectors, data@ 0
/dev/sdc: isw, "isw_dijhgahgdb", GROUP, ok, 1465149165 sectors, data@ 0
/dev/sdd: isw, "isw_dijhgahgdb", GROUP, ok, 1465149165 sectors, data@ 0

I'm playing around with some commands at the moment trying to get it set up properly. I'll put down whatever commands I use and hopefully it works out right.


Volume0 refers to the RAID volume name I set with the Intel controller on boot, isw makes it Intel Software raid. To set this in fstab my guess would be /dev/mapper/isw_dijhgahgdb_Volume0 has to be the source. If that is correct then the only issue I am still trying to figure out would be the partitioning of the array.

I tried another approach at setting up the array. I started over and used mdadm to create the arrays which went fine. I enabled logging for the array incase anything happened to it and right away I got sent an email saying that the array was degraded, same thing was being shown from the Intel manager the first attempt with it. I shut down and re-organized all the cables thinking something was loose, start back up and the same thing. Here is the output of the email.
Output of fdisk -l
By the looks of it /dev/sde is the problem. It has 1 more block than the other 3 identical drives?????? and somehow it formatted with ID ee instead of fd, and is EFI GPT file system instead of the Linux Raid that I told it to format with.
End of the day however so if there are any solutions I'll try them out in the morning.

Mmm the dmraid solution seemed to have defined the device. If you want to use mdadm then you would have to turn off all raid in the bios just making the disks ide disks. You can then use mdadm to setup software raid for the disks.

For the above command I would suggest doing

That should show more clearly what files are available after doing the


It seems to me at this point you had the raid device activated and working you just needed to partition this and apply a file system.

Yesterday I managed to get the array going in RAID10 using mdadm. As soon as I turned on the logging I got an email saying the array was degraded. Looks like one of the drives is no good anymore so it's getting RMA'd.