Hej all,
Sorry for the late answer..
.
.
UUID=19881aa7-699a-41ff-bd65-216e1d3de62c /var/lib/pgsql/9.5 xfs _netdev 0 0
[Unit]
Description=PostgreSQL 9.5 database server
After=syslog.target
After=network.target
After=remote-fs.target
.
.
.
3. systemctl daemon-reload
On Thu, Feb 18, 2016 at 8:32 AM, Alberto Cabello Sánchez <alberto@xxxxxxx> wrote:
On Wed, Feb 17, 2016 at 06:07:29PM +0100, proj@xxxxxxx wrote:
> Hi everybody,
>
> I installed a postgresql database on Redhat 7.1 and I decided to move the
> database on an ISCSI device (LUN) inside a logical volume, mounted at
> starting of the machine (xfs formatted). The mounting point is /var/lib/pgsql
>
> At the boot of the server, postgresql.service is in failed status.
>
> In messages.log :
> systemd: mounting /var/lib/pgsql
> starting PostgreSQL database server
> kernel sdv: unknown partition table
> sd 2:0:0:0: [sdb] attached SCSI disk
> xfs (dm-4): Mounting V4 Filesystem
> postgresql-check-db-dir: "/var/lib/pgsql/data" is missing or empty
> postgresql.service: control process exited, code=exited status=1
> Failed to start PostgreSQL database server.
>
>
> When I'm logged on the server, if it try to start manually the database :
> systemctl start postgresql --> OK (and I don't lose any data, database is
> available)
>
> I think it's a problem of order in the boot process : network service must
> be started, then iscsi, then lvm etc... So I tried to force dependencies
> on the /usr/lib/systemd/system/postgresql.service adding
> "After=lvm-pgscan.service iscsi.service" etc... but the result is
> the same : failure in starting postgresql
>
> Any ideas ?
Take a look at this thread:
http://www.gossamer-threads.com/lists/linux/kernel/1332888
I'm not an XFS guru, but it seems that XFS does a bunch of checks just
on/after mount, so perhaps PgSQL cannot access /var/lib/pgsql/data
immediately.
You could igive it a try forcing some delay in systemd conf file, or
tweaking the PgSQL startup script.
--
Alberto Cabello Sánchez
Universidad de Extremadura
