Building out more RDS functionality

fabfile.py

@@ -19,8 +19,8 @@ def update_prod():
     with cd("/opt/regluit"):
         run("./deploy/update-prod")
 
-def backup_db():
-    run("""TS=`date +"%Y-%m-%dT%H:%M:%S"`; /home/ubuntu/dump.sh | gzip > unglue.it.${TS}.sql.gz; scp ./unglue.it.${TS}.sql.gz  b235656@hanjin.dreamhost.com: ; rm -f unglue.it.${TS}.sql.gz""")
+def backup_db(name='unglue.it'):
+    run("""TS=`date +"%Y-%m-%dT%H:%M:%S"`; /home/ubuntu/dump.sh | gzip > {0}.${{TS}}.sql.gz; scp ./{0}.${{TS}}.sql.gz  b235656@hanjin.dreamhost.com: ; rm -f {0}.${{TS}}.sql.gz""".format(name))
 
 def get_dump():
     """Dump the current db on remote server and scp it over to local machine.
@@ -31,11 +31,23 @@ def get_dump():
     local("scp web1:/home/ubuntu/unglue.it.sql.gz .")
     local("gunzip -f unglue.it.sql.gz")
     
+def copy_dump_to_ry_dev():
+    """Dump the current db on remote server and scp it over to ry-dev.
+    Note:  web1 has been hardcoded here to represent the name of the unglue.it server
+    """
+    run("./dump.sh > unglue.it.sql ")
+    run("scp unglue.it.sql ry-dev.dyndns.org:")    
+    
+    
 def build_prod_instance(ami_id='ami-a29943cb'):
     """Build a new instance to serve as server instance for unglue.it"""
     # http://my.safaribooksonline.com/book/-/9781449308100/2dot-ec2-recipes/id2529379
     # default ami-a29943cb' is Ubuntu 12.04 Precise EBS boot
 
+def ecdsa():
+    """Calculate the host ECSDA host fingerprint http://bridge.grumpy-troll.org/2011/01/openssh.html """
+    run("""ssh-keygen -f /etc/ssh/ssh_host_ecdsa_key.pub -l""")
+    
 def ssh_fingerprint():
     """display ssh fingerprint of /home/ubuntu/.ssh/id_rsa.pub on remote machine"""
     run ("""ssh-keygen -l -f /home/ubuntu/.ssh/id_rsa.pub""")

sysadmin/aws.py

@@ -29,15 +29,26 @@ def all_rds():
 def all_rds_parameter_groups():
     return rds.get_all_dbparameter_groups()
     
-def modify_please1_pg_group():
-    """kinda ugly
+def modify_rds_parameter(group_name, parameter, value, apply_immediate=False):
+    """change parameter in RDS parameter group_name to value
     http://stackoverflow.com/a/9085381/7782
-    After doing this, I changed please db to talk to this parameter group and rebooted db
+    Remember to make sure that the parameter group is actually associated with the db.
+    You will likely need to reboot db too.
     """
-    pg = conn.get_all_dbparameters('mygroup')
-    pg2 = rds.get_all_dbparameters('mygroup', marker = pg.Marker)
-    pg2['tx_isolation'].value = True
-    pg2['tx_isolation'].apply(True)
+    
+    pg = rds.get_all_dbparameters(group_name)
+    while not pg.has_key(parameter) and hasattr(pg, 'Marker'):
+        pg = rds.get_all_dbparameters(group_name, marker = pg.Marker)
+    if pg.has_key(parameter):
+        pg[parameter].value = value
+        pg[parameter].apply(immediate=apply_immediate)
+        return True
+    else:
+        return False
+
+# how I associated the param_group to a db and then rebooted db
+# rds.modify_dbinstance(id='production', param_group='production1', apply_immediately=True)
+# rds.reboot_dbinstance(id='production')
     
 def all_snapshots(owner=GLUEJAR_ACCOUNT_ID):
     """by default, return only snapshots owned by Gluejar  -- None returns all snapshots available to us"""