Results 1 to 2 of 2
Enjoy an ad free experience by logging in. Not a member yet? Register.
- Join Date
- Jun 2011
Exchanging SATA controller when using software RAID0
Not so long ago I've changed my home Server/NAS machine to a new one. I have also decided to change the OS from Windows to Linux. I've chosen Ubuntu Server 11.04 (headless config). So you can say I'm a n00b.
First I've bought 2 very cheap PCI SATA controllers. And my disk config now is:
6x 500GB Seagate on motherboard integrated controller
4x 2TB WD on those cheap PCI controllers
My software RAID setup looks like this:
Personalities : [linear] [multipath] [raid0] [raid1] [raid6] [raid5] [raid4] [raid10]
md0 : active raid0 sde1 sdc1 sdd1 sdb1 sdf1 sda1
2894518272 blocks super 1.2 512k chunks
md1 : active raid0 sdc5 sdf5 sdb5 sdd5 sde5 sda5
35770368 blocks super 1.2 512k chunks
md2 : active raid0 sdj1 sdg1 sdh1 sdi1
7814049792 blocks super 1.2 4k chunks
md0 and md1 are located on disks connected to the integrated controller. md2 was created under webmin. I'm using ext4 (don't know if that is related to my questions).
My questions are about md2 (4 WD HDDs). I've bought a new controller (Adaptec AAR-1430SA) and I'm planning on exchanging those 2 cheap PCI controller to the new one.
1) Will I need to install something after adding the controller to my config ? (if so then what and how ?)
2) What to do to rebuild the RAID0 array after connecting those disks to the new controller ? (I don't want to loose any data - I know that I can't use the hardware solution provided by the controller)
Firstly, if you dont want to lose data, DO NOT USE RAID-0
if any one of the 6 drives in your raid-0 dies, you will likely lose all data on the raid! To avoid data loss, you'll likely want something like RAID-5 on the drives.
As for "rebuilding" the raid if your controller were to die for instance, all you'd need to do is replace the controller, and mdadm should pick them up automatically.
With software raid, the raid data itself, and metadata used to create the raid is stored on the drive itself. This means that you dont even need to worry about the order they get reconnected in. E.g. your sdb could be changed to sdc after, and tthe raid software (mdadm) will know this, and change accordingly."I am not an alcoholic, alcoholics go to meetings"
Registered Linux user = #372327