Thread: BUG #14451: PostgreSQL server on iscsi disks
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tim.klicks@gmail.com writes:
> As mentioned, I am not sure if this is a bug or not, but it would be great
> if you could include the remote-fs.target dependency into the
> postgresql-9.6.service file.

I'm fairly certain this was an intentional omission.  The normal
configuration does not put the database on a remote filesystem,
so adding this dependency would simply delay database startup
longer than necessary, and maybe even uselessly stop it from
coming up at all if there are problems on some remote filesystem.
The expectation is that if you rearrange things to store the DB
remotely, you should adjust the service file to match.  (Perhaps
that could be documented better, though.)

            regards, tom lane