Alvaro Herrera wrote:
> Bruce Momjian wrote:
>
>> Alvaro Herrera wrote:
>>
>
>
>>> Why not? Right now it's single-threaded. Would it be faster if it ran
>>> several copies in parallel?
>>>
>> Sure, but that assumes you have parallel I/O channels; I assume right
>> now it is I/O limited.
>>
>
> But so does parallel pg_restore, no?
>
>
Well, no, it assumes you're not saturating your I/O bandwidth, which
turns out to be the case more often than you might expect, even without
multiple I/O channels.
cheers
andrew