[execo_g5k] Fix check_deployed_command to account for miniOS or NVMe drives

This fixes two issues:

• while kadeploy is running, we get SSH access to the deployment miniOS. The current check_deployed_command was wrongly classifying the node as "deployed" in that case.

• Grid'5000 now has clusters with NVMe drives, and the current regexp was not handling that case (e.g. /dev/nvm0n1pX instead of /dev/sdbX). It would wrongly classify the node as "undeployed" even though it was already deployed.

This changes the logic to a positive one.

Note that I couldn't actually test this with execo. This is a kind of "backport" of the same change in Enoslib (that I tested):

discovery/enoslib!138 (merged)

src/execo_g5k/config.py

@@ -25,7 +25,7 @@ g5k_configuration = {
     'default_env_name': None,
     'default_env_file': None,
     'default_timeout': 900,
-    'check_deployed_command': "! (mount | grep -E '^/dev/[[:alpha:]]+2 on / ')",
+    'check_deployed_command': "mount | grep -q '^/dev/.*3 on / '",
     'no_ssh_for_local_frontend' : False,
     'polling_interval' : 20,
     'tiny_polling_interval' : 10,
@@ -61,7 +61,7 @@ g5k_configuration = {
   `execo_g5k.kadeploy.deploy` to check that the nodes are correctly
   deployed. This command should return 0 if the node is correctly
   deployed, or another value otherwise. The default checks that the
-  root is not on the second partition of the disk.
+  root filesystem is on the third partition of the disk.
 
 - ``no_ssh_for_local_frontend``: if True, don't use ssh to issue g5k
   commands for local site. If False, always use ssh, both for remote