PostgreSQL 9.4 – Changing Data Directory Location Using SSHFS

errorspostgresql-9.4remotestartup

I'm trying to change my PostgreSQL data directory initially on my Debian server to my NAS (running also on linux).
To do so, I've mounted it on a specific directory through sshfs (in /mnt/titanBDD).

Now I get this error when starting postgre (using service and pg_ctlcluster):

Error: could not exec /usr/lib/postgresql/9.4/bin/pg_ctl
/usr/lib/postgresql/9.4/bin/pg_ctl start -D
/var/lib/postgresql/9.4/redirect_TITAN -l
/var/log/postgresql/postgresql-9.4-main.log -s -o -c
config_file="/etc/postgresql/9.4/main/postgresql.conf"

Here is the modified part of postgresql.conf:

data_directory = '/var/lib/postgresql/9.4/main'

modified in:

data_directory = '/var/lib/postgresql/9.4/redirect_TITAN'

Where redirect_TITAN:

0 lrwxrwxrwx 1 root root 18 nov.   1 16:23 /var/lib/postgresql/9.4/redirect_TITAN -> /mnt/titanBDD/main

I've used the following command to move my data:

root@machine:~# cp --preserve=all -R /var/lib/postgresql/9.4/main/ /mnt/titanBDD/

Acutally, in /mnt/titanBDD I have:

root@machine:~# ls -als /mnt/titanBDD/
total 12
4 drwxr-xr-x 1 postgres postgres 4096 nov.   1 16:19 .
4 drwxr-xr-x 3 root     root     4096 nov.   1 14:05 ..
4 drwx------ 1 postgres postgres 4096 nov.   1 15:59 main

What is going wrong? I guess it's a context issue but don't have any idea where to search.

Best Answer

I finally solved the problem using -o allow_other when mounting the /mnt/titanBDD with sshfs.

The error message was still :

Error: could not exec

/usr/lib/postgresql/9.4/bin/pg_ctl /usr/lib/postgresql/9.4/bin/pg_ctl 
start -D /var/lib/postgresql/9.4/redirect_TITAN
-l /var/log/postgresql/postgresql-9.4-main.log -s -o -c
config_file="/etc/postgresql/9.4/main/postgresql.conf"

(which repeat the command twice?!)

when executing:

usr/lib/postgresql/9.4/bin/pg_ctl 
start -D /var/lib/postgresql/9.4/redirect_TITAN
-l /var/log/postgresql/postgresql-9.4-main.log -s -o -c
config_file="/etc/postgresql/9.4/main/postgresql.conf"

but I noticed the user postgres didn't have access to the mounted point which put me on the track.

Now PostgreSQL starts without any problem.

Related Question