oardocker start/stop but no reset

[npf]

could we make `oardocker stop' not reset to the install stage snapshot ?

-> add a `oardocker reset' command for that purpose, but otherwise make docker vm keep their states on stop, for the next start actions.

[SalemHarrache]

I started implementing it and it's a little bit more complicated than I expected.

• The file etc/hosts and ssh_config need to be updated because ip addresses change when we restart the containers ( :ok_hand: )
• Are init scripts (my_init.d) relaunched when doing a reset? Because error messages occur when these scripts are not idempotents. ( :question: )

E.g.

node1 | Adding user user1...
node1 | useradd: user 'user1' already exists

server | NOTICE:  CREATE TABLE will create implicit sequence "moldable_job_descriptions_moldable_id_seq1" for serial column "moldable_job_descriptions.moldable_id"
server | ERROR:  relation "moldable_job_descriptions" already exists
server | ERROR:  relation "job_mold" already exists
server | ERROR:  relation "log_mold_desc" already exists
server | ERROR:  relation "queues" already exists
server | ERROR:  relation "scheduler" already exists
...
server | DBD::Pg::db do failed: ERROR:  column "core" of relation "resources" already exists at /usr/local/lib/oar/oarproperty line 85.
server | DB error: ERROR:  column "core" of relation "resources" already exists

The question is: should we change scripts so that they work appropriately the second time or should we ignore my.init.d scripts when doing a oar reset?

[npf]

Ok. Let's keep that for later (or never) then.