On Wed, Jan 21, 2015 at 10:42 AM, Magnus Hagander <magnus@xxxxxxxxxxxx> wrote:
On Wed, Jan 21, 2015 at 7:19 PM, Matheus de Oliveira <matioli.matheus@xxxxxxxxx> wrote:On Wed, Jan 21, 2015 at 3:32 PM, Craig James <cjames@xxxxxxxxxxxxxx> wrote:# ls -l /data/postgres-9.3/main/pg_tblspc/16747lrwxrwxrwx 1 postgres postgres 27 2014-08-18 11:28 /data/postgres-9.3/main/pg_tblspc/16747 -> /postgres/tablespaces/uorsy/# du -sh /data/postgres-9.3/tablespaces/uorsy35G /data/postgres-9.3/tablespaces/uorsy# du -sh /data/postgres-9.3/tablespaces/uorsy/*35G /data/postgres-9.3/tablespaces/uorsy/82086248.1M /data/postgres-9.3/tablespaces/uorsy/PG_9.3_2013061214.0K /data/postgres-9.3/tablespaces/uorsy/pgsql_tmp4.0K /data/postgres-9.3/tablespaces/uorsy/PG_VERSION# find /data/postgres-9.3/tablespaces/uorsy \! -links 1 -type f | wc -l740Am I overlooking or is there something really wrong here?
First, all files of a tablespace should be inside PG_9.3_201306121 directory, why do you have those other files? Second, there shouldn't be any hard link inside of a tablespace, as PostgreSQL is not creating them, someone must have done it by hand. I'm guessing all inside PG_9.3_201306121 is linked to the root path of the tablespace, which is wrong.If I'm not overlooking, then neither barman nor pg_basebackup is to blame, but whoever created the hard links; if PostgreSQL did this (which I doubt) then it is a bug.Might it be hardlinks created by pg_upgrade? If so, they can just be removed...
OK, but which one? I'm pretty reluctant to do something that would destroy my entire database...
Craig
--
