I'm looking at an online man page for the sync command and I can't quite figure out the intended use of the -d
or --data
option.
Is it faster? Does it have any noticeable effect? Or is it something legacy?
Many filesystems have checksums for meta data. Usual data on the other side is often not checksummed. Using sync -d
is a faster operation than a complete sync
and should be preferred in situations where time matters, like hanging on a fail-safe battery. sync -d
also needs less IO access and thus can increase the lifespan of the device, especially on cheap devices like SDcards and other cheap flash memory technologies.
Recommend use:
Use sync -d
within program loops which are followed with a sync
at the end of loop and use sync
or sync -d; sync
in CLI.
sync -d
? There is no reference in the answer.