On 06/29/2017 12:05 AM, Ken Tanzer wrote:
> Thanks for the responses. For me, using the 9.2 binary was the winner.
> Shoulda thought of that!
>
> On Wed, Jun 28, 2017 at 1:30 PM, Tom Lane <tgl@sss.pgh.pa.us
> <mailto:tgl@sss.pgh.pa.us>> wrote:
>
>
> Generally speaking, it helps a lot if you don't insist on restoring the
> output in a single transaction. In this case, that would allow the
> restore to ignore the new parameters and move on.
>
> regards, tom lane
>
>
> Well sure, I can see it increases your chances of getting _something_
> restored. But there's also a lot to be said for ensuring that _all_
> your data restored, and did so correctly, no?
If you are using -l to pg_restore then you are also doing
--exit-on-error. In the case you showed(ERROR: unrecognized
configuration parameter "lock_timeout") that will not affect the data.
In fact in most cases that I have run across ERROR's are more
informational then data affecting.
>
> Cheers,
> Ken
>
>
> --
--
Adrian Klaver
adrian.klaver@aklaver.com