Hi,
Thanks for the report. Pushed 9.5.10-4 RPMs to RHEL 6 repo, which fixes this
issue. Apologies for the inconvenience.
Regards, Devrim
On Fri, 2017-12-08 at 19:15 +0000, Aaron Davis wrote:
> Hello folks,
>
> I'm finding that using the /etc/init.d/postgresql-9.5 script's initdb
> function is not work in the rpms obtained from postgresql.org's repo
>
> the steps to reproduce for me at least are to download:
> https://download.postgresql.org/pub/repos/yum/9.5/redhat/rhel-6.9-x86_64/post
> gresql95-server-9.5.10-3PGDG.rhel6.x86_64.rpm
> https://download.postgresql.org/pub/repos/yum/9.5/redhat/rhel-6.9-x86_64/post
> gresql95-libs-9.5.10-3PGDG.rhel6.x86_64.rpm
> https://download.postgresql.org/pub/repos/yum/9.5/redhat/rhel-6.9-x86_64/post
> gresql95-9.5.10-3PGDG.rhel6.x86_64.rpm
>
> then:
> rpm -i postgresql95-libs-9.5.10-3PGDG.rhel6.x86_64.rpm
> rpm -i postgresql95-9.5.10-3PGDG.rhel6.x86_64.rpm
> rpm -i postgresql95-server-9.5.10-3PGDG.rhel6.x86_64.rpm
>
> then run:
> service postgresql-9.5 initdb
> the response I get:
> Initializing database: [FAILED]
>
> I looks to me like the previous version of the script had this command:
> $SU -l postgres -c "$initdbcmd" >> "\$PGLOG\" 2>&1 &"
>
> While last night's update changed it to
>
> $SU -l postgres -c "$initdbcmd >> \"$PGLOG\" 2>&1 &"
>
> My shell scripting is kinda weak, but I believe the difference is the new
> iteration of the command pushes the '&' with the command allowing the su
> process that spawned it to return instantly killing the initdb process.
>
> Thank you for your time,
> Aaron Davis | Software Development, Plixer
> office 207 324 8805 /237
> aaron.davis@plixer.com<mailto:aaron.davis@plixer.com> | www.plixer.com<https:
> //www.plixer.com/>;
>
>
--
Devrim Gündüz
EnterpriseDB: https://www.enterprisedb.com
PostgreSQL Consultant, Red Hat Certified Engineer
Twitter: @DevrimGunduz , @DevrimGunduzTR
