Ennio-Sr <nasr.laili@tin.it> writes:
> * Tom Lane <tgl@sss.pgh.pa.us> [201112, 12:30]:
>> Use pg_controldata to print out the contents of the pg_control file as
>> extracted from the backup. If it doesn't say the cluster state is "shut
>> down", then you've probably got an inconsistent filesystem backup.
> Mmh... it does say 'shut down'; as to the other items to my
> un-experienced eyes they look so similar to the result I get on a bare
> 'new' (empty) cluster just created in a different dir...
In that case, in principle restoring the backup should work. Are you
sure you copied the entire $PGDATA directory tree? If you omitted
portions like pg_xlog or pg_clog that could be problematic. What
failure messages do you get *exactly*?
regards, tom lane