#601813 Please provide a more friendly/useful message instead of "Warning: The resulting partition is not properly aligned for best performance."

Package:
parted
Source:
parted
Description:
disk partition manipulator
Submitter:
Andras Korn
Date:
2014-12-14 06:39:09 UTC
Severity:
wishlist
#601813#5
Date:
2010-10-29 22:57:44 UTC
From:
To:
Hi,

I'm attempting to create a partition on a disk with a GPT.

I have a partition that ends at sector 501535; that was apparently fine
performance-wise, because parted printed no warning.

Now I would like to create a new partition, starting from sector 501536, but
parted tells me that's not optimal. It doesn't tell me what _would_ be
optimal, even though it could.

Seeing that I started parted with -a optimal, which supposedly automatically
aligns partitions optimally, it'd be great if it could at least tell me what
the nearest "optimal" sector to the one I chose is.

The only workaround I found was to switch to some unit larger than sectors,
which resulted in the beginning of the new partition to be shifted forward
by 224.

Please either print the suggested "optimal" value or even offer to use that
instead of what the user supplied.

Thanks.

Andras

#601813#12
Date:
2014-12-14 05:36:35 UTC
From:
To:
found 601813 2.3-12
stop

This is still present in newer versions.  In interactive mode you get the
message:

"Warning: The resulting partition is not properly aligned for best performance.
Ignore/Cancel?"

Reading #643625, this message is given for any disk slice not aligned to
1MB.  Given the circumstances it appears in, the message is about as
useful as displaying a gigantic question mark on screen since this leaves
no clue this is based on arbitrary (though not entirely unreasonable)
built-in limit which is neither documented nor findable by looking up
information on the disk drive.

In addition to Andras Korn's suggestion of printing the "optimal" values,
I'd suggest "align", "+align" and "-align" options for the above prompt.
These would be for nearest "optimal" alignment, nearest higher "optimal"
sector, and nearest lower "optimal" sector.

I must also point out it is quite possible 1MB may not be too small to be
optimal with future storage devices.  I suspect the technology behind
Western Digital/HGST's "shingled magnetic recording" may well result in
an optimal write size distinctly larger than 1MB (and Flash could
follow).

I'm inclined to rate this as distinctly worse than wishlist given how
often people will run into this and how bad the behavior is.