On 03/07/17 14:34, Carlos E. R. wrote:
On 2017-07-03 08:10, David C. Rankin wrote:
See:
Not from the horse's mouth, I notice ... :-) (Yes I do know Arch has a good reputation :-)
Create the Partition Table (GPT)
It is highly recommended to pre-partition the disks to be used in the array. Since most RAID users are selecting HDDs >2 TB, GPT partition tables are required and recommended. Disks are easily partitioned using gptfdisk.
After created, the partition type should be assigned hex code FD00.
I think you will find this is cargo cult monkey see monkey do :-)
If a larger disk array is employed, consider assigning disk labels or partition labels to make it easier to identify an individual disk later. Creating partitions that are of the same size on each of the devices is preferred. A good tip is to leave approx 100 MB at the end of the device when partitioning. See below for rationale.
Huh?
Let's see:
«When replacing a failed disk of a RAID, the new disk has to be exactly the same size as the failed disk or bigger — otherwise the array recreation process will not work. Even hard drives of the same manufacturer and model can have small size differences. By leaving a little space at the end of the disk unallocated one can compensate for the size differences between drives, which makes choosing a replacement drive model easier. Therefore, it is good practice to leave about 100 MB of unallocated space at the end of the disk.»
Curious!
This is real, and does cause problems ... a 3TB disk (for example) is guaranteed to be *at least* three million million bytes. But depending on manufacturing (now that Cylinders, Heads, Sectors is just a fiction) there's no constraint on how much (or little) over that figure is acceptable. So disks do vary ... a case occurred on the raid list maybe six months ago?
It also says:
«Note: It is also possible to create a RAID directly on the raw disks (without partitions), but not recommended because it can cause problems when swapping a failed disk.»
Given that the guy who WROTE a large chunk of the raid code uses raw disks as a matter of course, I'm afraid I wouldn't give much credence to that. Where it does matter, is you cannot *boot* off a partition on a raw raid disk. Pretty much all boot code relies on an MBR or GPT to locate the OS. The other reason for not using raw disks is a lot of disk tools assume "no gpt/mbr == blank disk". You don't want an install CD to stomp all over your raid because it assumed you weren't using that disk. All this has been discussed "recently" on the linux raid list, and I could probably find it in my archive, but with three years to search that's a lot of emails ... :-) Cheers, Wol -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org