[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: formatting a 12.73 Tb disk on ROCKS 5.2



m roth 5-cent us wrote:
> Doll, Margaret Ann wrote:
>> On Thu, May 9, 2013 at 3:09 PM, <m roth 5-cent us> wrote:
>>
>>> Doll, Margaret Ann wrote:
>>> > The answer came from my vendor at Atipa.
>>> >
>>> > You will need to change the partitioning scheme to use GPT in order
>>> to
>>> > enable greater than 2TB support.
>>>
>>> Which is what I was saying.
>>> >
>>> > # parted -s /dev/sda mklabel gpt
>>>
>>> This is command-line version of what I said - I was giving you the
>>> interactive version. I do not understand why the above would work, and
>>> what I suggested you do gave you "invalid token".
>>> >
>>> > # parted –s /dev/sda rm 1
>>>
>>> The above removes partition 1.
>>> >
>>> > # parted –s /dev/sda “mkpart primary xfs 1 -1”
>>> >
>>> You've decided to use xfs, yes? And the 1 bothers me, a lot. That's
>>> either sector or cylinder... and they did *not* tell you to use -a
> optimal for
>>> aligning the partition. If you do parted -l, what do you see?
>>>
>> [root nas-0-0 ~]# parted -l
>>
>> Model: AMCC 9650SE-16M DISK (scsi)
>> Disk /dev/sda: 14.0TB
>> Sector size (logical/physical): 512B/512B
>> Partition Table: gpt
>>
>> Number  Start   End     Size    File system  Name     Flags
>>  1      17.4kB  14.0TB  14.0TB  xfs          primary
>>
>> df -h
>>
>> /dev/sda1              13T  8.4G   13T   1% /bigdisk1
>
> Yup - if this were a single disk, that would not be a good place to start,
> esp. if it was a large disk. I don't know how big the drives in the array
> are, and even though parted says it's 512/512 sector size, most new large
> drives are, in reality, 4k in hardware/firmware. You *might* consider
> repartitioning, but start at 2048k, rather than 1. I suspect that if you
> did
>       parted align-check optimal /dev/sda1
> that it would tell you it was not aligned properly.
>
Hit <send> too soon.... The reason alignment matters is that it affects
throughput speed.

       mark



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]