Merge pull request #12 from VISIONAI/tagfix

Tagfix

README.md

@@ -3,11 +3,10 @@
     VMX makes computer vision easy.  Docker makes bundling complex
     applications easy.  The vmx-docker-manager lets you run the VMX
     Object Detection server on a Linux computer, or a (Mac/Windows)
-    machine via boot2docker. [We also provide a non-docker native Mac
-    OS X build of VMX].
+    machine via boot2docker. 
 
 ##Requirements
-    64 bit Linux and Docker >= 1.2
+    64 bit Linux and Docker >= 1.2, root permissions to run Docker
     boot2docker for Mac or Windows
 
 ##Instructions
@@ -23,14 +22,14 @@ Or download the shell script directly: https://raw.githubusercontent.com/VISIONA
     Commands:
         start PORT: Start vmx on localhost:PORT
         stop:       Stop vmx
-        enter:      Shell access to docker
+        enter:      Shell access to a Docker container for inspecting volumes
         init :      Download and start the containers that have the pre-reqs, mcr, etc.
                     (This is done automatically the first time you start vmx)
         update TAG: Update your vmx (will stop vmx first). TAG must be
-                    "latest" or "dev", defaults to "latest"
-        backup BACKUP: Backup your vmx data (models, session logs,
-                       config file) to folder BACKUP
-        
+                    "latest" or "dev", defaults to "latest" when TAG is empty
+        backup BACKUP: Backup your vmx data (models and config file) to folder BACKUP
+        restore BACKUP: Restore from a backup        
+
 ##Example:
  Run vmx on port 3000
 
@@ -61,7 +60,17 @@ or
     $  ./vmx update dev
 
 
+##Permissions
+
+vmx-docker-manager uses Docker which requires permission to run as root, or the
+user to be added to the `docker` group.
+
+To make sure Docker is installed and running correctly on your machine, you can first
+try a simple Docker command such as:
 
+```
+docker run -t -i --rm ubuntu echo "Docker runs"
+```
 
 ##Explanation
 
@@ -73,9 +82,13 @@ general, VMX requires a few files/folder to be available:
    - Local Model storage
  - Certain open source libraries
 
-The VMX Docker manager is a set of sane defaults and configurations that store user data (session and models) in seperate volumes at /vmx/sessions and /vmx/models; and loads the binaries into /vmx/build.  Those mount points, along with the matlab MCR dependency are run within the context of an Ubuntu 14.04 with the required libraries.
+The VMX Docker manager is a set of sane defaults and configurations
+that store user data (session and models) in seperate volumes at
+/vmx/sessions and /vmx/models; and loads the binaries into /vmx/build.
+Those mount points, along with the matlab MCR dependency are run
+within the context of an Ubuntu 14.04 with the required libraries.
 
-The dockerfiles that build the referenced docker images can be found here:
+The Dockerfiles that build the referenced Docker images can be found here:
 
 - https://github.com/VISIONAI/docker-mcr-2014a
 - https://github.com/VISIONAI/docker-vmx-userdata

vmx

@@ -1,7 +1,7 @@
 #!/bin/sh
 # VMX Docker Manager by vision.ai
 # Starts the vision.ai object recognition server
-# Requires docker >= 1.2; and perms to run docker
+# Requires docker >= 1.2; and permissions to run docker
 # Visit us on the web: 
 #   https://vision.ai
 # Github repository for this file:
@@ -19,23 +19,20 @@ FOLDERNAME=`date +%Y%a%b%d_%H%M%S`
 
 init () {
   echo "Starting containers..."
-  if [ "$VMXTAG" = ""]; then
-      VMXTAG=":$VMXTAG"
-  fi
 
   docker run -d --name vmx-mcr visionai/mcr-2014a 2>/dev/null
   docker run -d --name vmx-userdata visionai/vmx-userdata 2>/dev/null
-  docker run -d --name vmx-middle:$VMXTAG visionai/vmx-middle 2>/dev/null
-  docker run -d --name vmx-server:$VMXTAG visionai/vmx-server 2>/dev/null
-  docker run -d --name vmx-appbuilder:$VMXTAG visionai/vmx-appbuilder 2>/dev/null
+  docker run -d --name vmx-middle$VMXTAG visionai/vmx-middle 2>/dev/null
+  docker run -d --name vmx-server$VMXTAG visionai/vmx-server 2>/dev/null
+  docker run -d --name vmx-appbuilder$VMXTAG visionai/vmx-appbuilder 2>/dev/null
   touch .vmx-init
 }
 
 
 usage () {
   echo "VMX Docker Manager by vision.ai"
   echo "Version: " `git describe --tags --dirty 2>/dev/null || echo no-git`
-  echo "Requires docker >= 1.2; and perms to run docker"
+  echo "Requires docker >= 1.2; and permissions to run docker"
   echo "==============================================="
   echo "Usage: vmx COMMAND [OPTIONS]"
   echo 
@@ -236,15 +233,15 @@ update_vmx(){
   docker rm vmx-server 2>/dev/null
   docker rm vmx-appbuilder 2>/dev/null
 
-  docker pull visionai/vmx-middle:$VMXTAG
-  docker pull visionai/vmx-server:$VMXTAG
-  docker pull visionai/vmx-appbuilder:$VMXTAG
+  docker pull visionai/vmx-middle$VMXTAG
+  docker pull visionai/vmx-server$VMXTAG
+  docker pull visionai/vmx-appbuilder$VMXTAG
 
-  docker run --name vmx-middle visionai/vmx-middle:$VMXTAG
+  docker run --name vmx-middle visionai/vmx-middle$VMXTAG
   A=$?
-  docker run --name vmx-server visionai/vmx-server:$VMXTAG
+  docker run --name vmx-server visionai/vmx-server$VMXTAG
   B=$?
-  docker run --name vmx-appbuilder visionai/vmx-appbuilder:$VMXTAG
+  docker run --name vmx-appbuilder visionai/vmx-appbuilder$VMXTAG
   C=$?
 
   #make a temp container for copying
@@ -281,7 +278,6 @@ case "$cmd" in
       init
     fi
     export VMXHOSTPORT=$2
-    export VMXTAG=$3
     start_vmx
     exit 0
     ;;
@@ -293,7 +289,13 @@ case "$cmd" in
     ;;
 
   update)
+
+    #Keep an empty TAG empy, and turn "dev" into ":dev"
     export VMXTAG=$2
+    if [ "$VMXTAG" != "" ]; then
+      VMXTAG=":$VMXTAG"
+    fi
+
     update_vmx
     init
     echo "Updated vmx"
@@ -318,10 +320,12 @@ case "$cmd" in
       if [ ! -d .vmx-backup ]; then
           echo "No .vmx-backup found.. have you done any backups?"
       else
-	 echo "try one of "
+	 echo "Try one of the following commands:"
 	 for b in .vmx-backup/*
          do
-	 echo ./vmx restore $b "\n"
+	     echo ./vmx restore $b
+             NMODELS=`ls $b/models 2>/dev/null | wc -l`
+             echo "Has ${NMODELS} models\n"
          done
 
          exit 1