[build_templates] [hostcfgd] Keep containers hostname up to date (#2924)

* Add updateHostName function to docker_image_ctl.j2
* Add hostname specification on container creating step
* Add listener for hostname changes in hostcfgd

Signed-off-by: Myron Sosyak <msosyak@barefootnetworks.com>

files/build_templates/docker_image_ctl.j2

@@ -5,6 +5,32 @@ function getMountPoint()
     echo $1 | python -c "import sys, json, os; mnts = [x for x in json.load(sys.stdin)[0]['Mounts'] if x['Destination'] == '/usr/share/sonic/hwsku']; print '' if len(mnts) == 0 else os.path.basename(mnts[0]['Source'])" 2>/dev/null
 }
 
+function updateHostName()
+{
+    HOSTS=/etc/hosts
+    HOSTS_TMP=/etc/hosts.tmp
+
+    EXEC="docker exec -i {{docker_container_name}} bash -c"
+
+    NEW_HOSTNAME="$1"
+    HOSTNAME=`$EXEC "hostname"`
+    if ! [[ $HOSTNAME =~ ^[a-zA-Z0-9.\-]*$ ]]; then
+        HOSTNAME=`hostname`
+    fi
+
+    # copy HOSTS to HOSTS_TMP
+    $EXEC "cp $HOSTS $HOSTS_TMP"
+    # remove entry with hostname
+    $EXEC "sed -i \"/$HOSTNAME$/d\" $HOSTS_TMP"
+    # add entry with new hostname
+    $EXEC "echo -e \"127.0.0.1\t$NEW_HOSTNAME\" >> $HOSTS_TMP"
+
+    echo "Set hostname in {{docker_container_name}} container"
+    $EXEC "hostname '$NEW_HOSTNAME'"
+    $EXEC "cat $HOSTS_TMP > $HOSTS"
+    $EXEC "rm -f $HOSTS_TMP"
+}
+
 function getBootType()
 {
     local BOOT_TYPE
@@ -105,6 +131,10 @@ start() {
     # Obtain our HWSKU as we will mount directories with these names in each docker
     HWSKU=`sonic-cfggen -d -v 'DEVICE_METADATA["localhost"]["hwsku"]'`
     {%- endif %}
+    HOSTNAME=`sonic-cfggen -m -v 'DEVICE_METADATA["localhost"]["hostname"]'`
+    if [ -z "$HOSTNAME" ] || ! [[ $HOSTNAME =~ ^[a-zA-Z0-9.\-]*$ ]]; then
+        HOSTNAME=`hostname`
+    fi
 
     DOCKERCHECK=`docker inspect --type container {{docker_container_name}} 2>/dev/null`
     if [ "$?" -eq "0" ]; then
@@ -121,6 +151,7 @@ start() {
             {%- endif %}
             preStartAction
             docker start {{docker_container_name}}
+            updateHostName "$HOSTNAME"
             postStartAction
             exit $?
         fi
@@ -167,6 +198,7 @@ start() {
         --tmpfs /tmp \
 {%- endif %}
         --tmpfs /var/tmp \
+        --hostname "$HOSTNAME" \
         --name={{docker_container_name}} {{docker_image_name}}:latest || {
             echo "Failed to docker run" >&1
             exit 4
@@ -186,11 +218,13 @@ stop() {
 }
 
 case "$1" in
-    start|wait|stop)
+    start|wait|stop|updateHostName)
-        $1
+        cmd=$1
+        shift
+        $cmd $@
         ;;
     *)
-        echo "Usage: $0 {start|wait|stop}"
+        echo "Usage: $0 {start|wait|stop|updateHostName new_hostname}"
         exit 1
         ;;
 esac

files/image_config/hostcfgd/hostcfgd

@@ -2,6 +2,7 @@
 # -*- coding: utf-8 -*-
 
 import os
+import re
 import sys
 import subprocess
 import syslog
@@ -22,6 +23,15 @@ TACPLUS_SERVER_TIMEOUT_DEFAULT = "5"
 TACPLUS_SERVER_AUTH_TYPE_DEFAULT = "pap"
 
 
+def is_valid_hostname(name):
+    if hostname[-1] == ".":
+        hostname = hostname[:-1] # strip exactly one dot from the right, if present
+    if len(hostname) > 253:
+        return False
+    allowed = re.compile("(?!-)[A-Z\d-]{1,63}(?<!-)$", re.IGNORECASE)
+    return all(allowed.match(x) for x in hostname.split("."))
+
+
 def is_true(val):
     if val == 'True' or val == 'true':
         return True
@@ -148,6 +158,7 @@ class HostConfigDaemon:
         tacacs_server = self.config_db.get_table('TACPLUS_SERVER')
         self.aaacfg = AaaCfg()
         self.aaacfg.load(aaa, tacacs_global, tacacs_server)
+        self.hostname_cache=""
 
     def aaa_handler(self, key, data):
         self.aaacfg.aaa_update(key, data)
@@ -166,10 +177,49 @@ class HostConfigDaemon:
             log_data['passkey'] = obfuscate(log_data['passkey'])
         syslog.syslog(syslog.LOG_INFO, 'value of {} changed to {}'.format(key, log_data))
 
+    def hostname_handler(self, key, data):
+        if key != "localhost":
+            return
+
+        hostname = data.get("hostname")
+
+        if not hostname:
+            syslog.syslog(syslog.LOG_WARNING, "hostname key is missing")
+            return
+        if not is_valid_hostname(hostname):
+            return
+        if hostname == self.hostname_cache:
+            return
+
+        syslog.syslog(syslog.LOG_INFO, "Get all running containers")
+        cmd = 'docker ps --format "{{.Names}}"'
+        try:
+            containers = subprocess.check_output(cmd, shell=True).split("\n")[:-1]
+        except subprocess.CalledProcessError as err:
+            syslog.syslog(syslog.LOG_ERR, "{} - failed: return code - {}, output:\n{}"
+                          .format(err.cmd, err.returncode, err.output))
+
+        for name in containers:
+            script = '/usr/bin/{}.sh'.format(name)
+            exists = os.path.isfile(script)
+            if not exists:
+                syslog.syslog(syslog.LOG_ERR, "Can't find control script for {}".format(name))
+                continue
+
+            cmd = "{} updateHostName {}".format(script, hostname)
+            try:
+                subprocess.check_call(cmd, shell=True)
+            except subprocess.CalledProcessError as err:
+                syslog.syslog(syslog.LOG_ERR, "{} - failed: return code - {}, output:\n{}"
+                              .format(err.cmd, err.returncode, err.output))
+
+        self.hostname_cache = hostname
+
     def start(self):
         self.config_db.subscribe('AAA', lambda table, key, data: self.aaa_handler(key, data))
         self.config_db.subscribe('TACPLUS_SERVER', lambda table, key, data: self.tacacs_server_handler(key, data))
         self.config_db.subscribe('TACPLUS', lambda table, key, data: self.tacacs_global_handler(key, data))
+        self.config_db.subscribe('DEVICE_METADATA', lambda table, key, data: self.hostname_handler(key, data))
         self.config_db.listen()