Initializing a disk from with GPT

Mayu Inc. pomomayu at gmail.com
Tue Jan 31 17:18:15 PST 2023


Hey like mentioned it seems like the disk maybe got damaged from trimming
hah. Initialization with GPT worked on another SSD that wasn't trimmed (I
think) - also a Samsung 860 EVO. We successfully installed a system with
pseudo filesystems.

Interestingly fdisk worked on that previous disk that rejected gpt. We were
able to test it on a throw away laptop. Just wanted to see if it would work
out of curiosity and it did allow for a successful install. Got some weird
warnings but won't include those in this thread.

Anyways these are the steps that were used:

# dd if=/dev/urandom of=/dev/da0 bs=32k status=progress
# gpt init -f -B da0
# gpt show da0
# disklabel64 -e da0s1

Proceed with manual installation.

On Sun, Jan 29, 2023 at 10:43 AM Mayu Inc. <pomomayu at gmail.com> wrote:

> Awesome thank you.
>
> Yes I chose the -E to opt for trimming. Running the command 'gpt init -f
> -E' without the -B did not cause the machine to freeze - it only froze with
> the -B flag.
>
> Actually not sure if 'froze' is a fair description. It just sat there with
> a cursor and would not take any additional commands.
>
> However after restarting the machine and running 'gpt show da0' everything
> seemed to be intact.
>
> On Sun, Jan 29, 2023 at 6:36 PM Matthew Dillon <dillon at backplane.com>
> wrote:
>
>> There is no disklabel on s0, it contains a msdos filesystem directly on
>> slice 0 ... that's what the BIOS EFI boot code accesses.  The only thing on
>> that msdos filesystem is the EFI boot binary.
>>
>> The disklabel on s1 would contain 'a' and 'd' (and usually a 'b' swap
>> partition too).
>>
>> I'll look into the -E issue.  It could be the SSD and not the system...
>> hardware is not always happy getting issued thousands of SATA TRIM commands
>> to clean the drive out.
>>
>> -Matt
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.dragonflybsd.org/pipermail/users/attachments/20230131/915d191e/attachment-0002.htm>


More information about the Users mailing list