Merge pull request #121 from isard-vdi/release/1.1.0

Release/1.1.0

CHANGELOG.md

@@ -2,9 +2,32 @@
 
 All notable changes to this project will be documented in this file.
 
-## [1.1.0-rc1] - 2019-02-03
+## [1.1.0] - 2019-02-26 | Canigó
 
-This is the first release candidate for the 1.1.0 version. The changelog is going to be updated after the release (approximately in a week [2019-02-10]). Check the [pull request](https://github.com/isard-vdi/isard/pull/94) for more information.
+### Added
+- Support under VMWare [#98](https://github.com/isard-vdi/isard/issues/98)
+- Support on AMD cpu [#98](https://github.com/isard-vdi/isard/issues/98)
+- Permissions on templates and bases in admin mode [#118](https://github.com/isard-vdi/isard/issues/118)
+- Delete templates and bases with all the derived desktops and templates (admin mode) [#114](https://github.com/isard-vdi/isard/issues/114)
+- Complete domain chain in domain dictionary [#78](https://github.com/isard-vdi/isard/issues/78)
+- Delete media will delete the media in all domains [#111](https://github.com/isard-vdi/isard/issues/111)
+
+### Changed
+- Force host-passthrough cpu mode in domains for better compatibility [#101](https://github.com/isard-vdi/isard/issues/101)
+- Swapped wizard steps hypervisor and engine as makes more sense to check first the engine. [#81](https://github.com/isard-vdi/isard/issues/81)
+- Domain status engine detection now done using stats thread. Not relaying only in libvirt events. [#120](https://github.com/isard-vdi/isard/issues/120)
+
+### Fixed
+- On physical host reboot the hypervisor docker gets online again. [#119](https://github.com/isard-vdi/isard/issues/119)
+- Media status correctly shown in web interface. [#110](https://github.com/isard-vdi/isard/issues/110)
+- Post installation updates register now works. [#109](https://github.com/isard-vdi/isard/issues/109)
+- Admin base and template modals not shown. [#112](https://github.com/isard-vdi/isard/issues/112)
+- Restart download when failed [#91](https://github.com/isard-vdi/isard/issues/91)
+- Delete process is now more atomic and will delete domain from database even if there are problems during disk delete. [#117](https://github.com/isard-vdi/isard/issues/117)
+
+### Removed
+- Windows install checkbox on creating domain. Not needed anymore. [#115](https://github.com/isard-vdi/isard/issues/115)
+- Global actions removed from templates and bases as are not needed there. [#116](https://github.com/isard-vdi/isard/issues/116)
 
 ## [1.0.1] - 2018-12-27
 

README.md

@@ -14,6 +14,7 @@ Open Source KVM Virtual Desktops based on KVM Linux and dockers.
 Pull images and bring it up:
 
 ```
+wget https://raw.githubusercontent.com/isard-vdi/isard/master/docker-compose.yml
 docker-compose pull
 docker-compose up -d
 ```

docker-compose.yml

@@ -23,7 +23,7 @@ services:
       - "443:443"
     networks:
       - isard_network
-    image: isard/nginx:1.1.0-rc1
+    image: isard/nginx:1
     restart: unless-stopped
     depends_on:
       - isard-app
@@ -36,11 +36,11 @@ services:
       - "/opt/isard/certs/default:/etc/pki/libvirt-spice"
       - "/etc/localtime:/etc/localtime:ro"
     ports:
-      - "5900-5949:5900-5949"
-      - "55900-55949:55900-55949"
+      - "5900-5999:5900-5999"
+      - "55900-55999:55900-55999"
     networks:
       - isard_network
-    image: isard/hypervisor:1.1.0-rc1
+    image: isard/hypervisor:1
     privileged: true
     restart: unless-stopped
 
@@ -58,7 +58,7 @@ services:
       - "isard-engine:127.0.0.1"
     networks:
       - isard_network
-    image: isard/app:1.1.0-rc1
+    image: isard/app:1
     restart: unless-stopped
     depends_on:
       - isard-database

dockers/app/certs.sh

@@ -1,5 +1,20 @@
 #!/bin/bash
 
+echo "Checking for isard-hypervisor ssh..."
+
+i=0
+while ! nc -z isard-hypervisor 22; do   
+  sleep 0.5
+  ((i++))
+  if [[ "$i" == '25' ]]; then
+    break
+  fi  
+  echo "Checking for isard-hypervisor shh"
+done
+
+echo "Adding isard-hypervisor keys"
+ssh-keyscan -T 10 isard-hypervisor > /root/.ssh/known_hosts
+
 # Remove all isard-hypervisor lines from known_hosts
 sed -i '/isard-hypervisor/d' /root/.ssh/known_hosts
 

extras/app-devel/devel-debug.yml

@@ -9,6 +9,7 @@ services:
     ##### - only devel - ############################
     ports:
       - "8080:8080"
+      - "28015:28015"
     expose:
       - "28015"
     #################################################

src/engine/controllers/broom.py

@@ -85,7 +85,7 @@ def polling(self):
                 domain_id = d['id']
                 status = d['status']
                 hyp_started = d['hyp_started']
-                if hyp_started is bool:
+                if type(hyp_started) is bool:
                     continue
                 if len(hyp_started) == 0:
                     continue
@@ -104,7 +104,7 @@ def polling(self):
                             if domain_id in hyps_domain_started[hyp_started]['active_domains']:
                                 logs.broom.debug('DOMAIN: {} ACTIVE IN HYPERVISOR: {}'.format(domain_id, hyp_started))
                                 state_libvirt = hyps_domain_started[hyp_started]['hyp'].domains[domain_id].state()
-                                state_str, cuase = state_and_cause_to_str(state_libvirt[0], state_libvirt[1])
+                                state_str, cause = state_and_cause_to_str(state_libvirt[0], state_libvirt[1])
                                 status = dict_domain_libvirt_state_to_isard_state(state_str)
                                 logs.broom.debug(
                                         'DOMAIN: {} ACTIVE IN HYPERVISOR: {} WITH STATUS: {}'.format(domain_id,

src/engine/controllers/ui_actions.py

@@ -27,6 +27,7 @@
     get_host_disk_operations_from_path, create_cmd_disk_from_scratch
 from engine.services.log import *
 
+DEFAULT_HOST_MODE = 'host-passthrough'
 
 class UiActions(object):
     def __init__(self, manager):
@@ -49,7 +50,7 @@ def start_domain_from_id(self, id, ssl=True):
 
         id_domain = id
         pool_id = get_pool_from_domain(id_domain)
-        cpu_host_model = self.manager.pools[pool_id].conf.get('cpu_host_model','host-model')
+        cpu_host_model = self.manager.pools[pool_id].conf.get('cpu_host_model',DEFAULT_HOST_MODE)
 
         # TODO: Read the cpu_host_model value from hypervisor_pool database
         try:
@@ -82,6 +83,12 @@ def start_paused_domain_from_xml(self, xml, id_domain, pool_id):
                 #     dict_action['start_after_created'] = True
                 #else:
 
+                if LOG_LEVEL == 'DEBUG':
+                    print(f'%%%% DOMAIN CREATING:{id_domain} -- XML TO START PAUSED IN HYPERVISOR {next_hyp} %%%%')
+                    print(xml)
+                    update_table_field('domains', id_domain, 'xml_to_start', xml)
+                    print('%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%')
+
                 self.manager.q.workers[next_hyp].put(
                     dict_action)
                 update_domain_status(status='CreatingDomain',
@@ -132,6 +139,12 @@ def start_domain_from_xml(self, xml, id_domain, pool_id='default'):
                 #                      detail='desktop starting paused in pool {} on hypervisor {}'.format(pool_id,
                 #                                                                                          next_hyp))
 
+                if LOG_LEVEL == 'DEBUG':
+                    print(f'%%%% DOMAIN: {id_domain} -- XML TO START IN HYPERVISOR: {next_hyp} %%%%')
+                    print(xml)
+                    update_table_field('domains',id_domain,'xml_to_start',xml)
+                    print('%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%')
+
                 self.manager.q.workers[next_hyp].put({'type': 'start_domain', 'xml': xml, 'id_domain': id_domain})
             else:
                 log.error('get next hypervisor in pool {} failed'.format(pool_id))
@@ -254,8 +267,7 @@ def deleting_disks_from_domain(self, id_domain, force=False):
         dict_domain = get_domain(id_domain)
 
         if dict_domain['kind'] != 'desktop' and force is False:
-            log.error('{} is a template, disks can not be deleted')
-            return -1
+            log.info('{} is a template, disks will be deleted')
 
         if len(dict_domain['hardware']['disks']) > 0:
             index_disk = 0
@@ -678,7 +690,7 @@ def creating_disks_from_template(self,
                     'Creating disk operation failed when insert action in queue for disk operations. Exception: {}'.format(
                         e))
 
-    def updating_from_create_dict(self, id_domain):
+    def updating_from_create_dict(self, id_domain,ssl=True):
         try:
             populate_dict_hardware_from_create_dict(id_domain)
         except Exception as e:
@@ -714,7 +726,8 @@ def updating_from_create_dict(self, id_domain):
 
         kind = get_domain_kind(id_domain)
         if kind == 'desktop':
-            xml_to_test = recreate_xml_to_start(id_domain)
+            cpu_host_model = self.manager.pools[pool_id].conf.get('cpu_host_model', DEFAULT_HOST_MODE)
+            xml_to_test = recreate_xml_to_start(id_domain,ssl,cpu_host_model)
             self.start_paused_domain_from_xml(xml=xml_to_test, id_domain=id_domain, pool_id=pool_id)
         else:
             update_domain_status('Stopped', id_domain,
@@ -724,7 +737,7 @@ def updating_from_create_dict(self, id_domain):
 
     def creating_and_test_xml_start(self, id_domain, creating_from_create_dict=False,
                                     xml_from_virt_install=False,
-                                    xml_string=None):
+                                    xml_string=None,ssl=True):
         if creating_from_create_dict is True:
             try:
                 populate_dict_hardware_from_create_dict(id_domain)
@@ -790,7 +803,8 @@ def creating_and_test_xml_start(self, id_domain, creating_from_create_dict=False
         else:
             #change viewer password, remove selinux options and recreate network interfaces
             try:
-                xml = recreate_xml_to_start(id_domain)
+                cpu_host_model = self.manager.pools[pool_id].conf.get('cpu_host_model', DEFAULT_HOST_MODE)
+                xml = recreate_xml_to_start(id_domain,ssl,cpu_host_model)
             except Exception as e:
                 log.error('recreate_xml_to_start in domain {}'.format(id_domain))
                 log.error('Traceback: \n .{}'.format(traceback.format_exc()))

src/engine/models/domain_xml.py

@@ -502,7 +502,7 @@ def set_cpu_host_model(self, cpu_host_model='host-model'):
             cpu.append(fallback)
 
         elif cpu_host_model == 'host-passthrough':
-            cpu = etree.parse(StringIO("cpu mode='{}' > </cpu>".format(cpu_host_model))).getroot()
+            cpu = etree.parse(StringIO("<cpu mode='{}' > </cpu>".format(cpu_host_model))).getroot()
             cpu.append(fallback)
 
         elif cpu_host_model in CPU_MODEL_NAMES:

src/engine/models/hyp.py

@@ -17,6 +17,7 @@
 from io import StringIO
 from collections import deque
 from statistics import mean
+import time
 
 import libvirt
 import paramiko
@@ -29,7 +30,8 @@
 from engine.services.db import get_id_hyp_from_uri, update_actual_stats_hyp, update_actual_stats_domain
 from engine.services.log import *
 from engine.config import *
-
+from engine.services.lib.functions import exec_remote_cmd
+from engine.services.db.domains import update_domain_status, get_domains_with_status_in_list
 
 TIMEOUT_QUEUE = 20
 TIMEOUT_CONN_HYPERVISOR = 4 #int(CONFIG_DICT['HYPERVISORS']['timeout_conn_hypervisor'])
@@ -48,7 +50,7 @@
 HYP_STATUS_ERROR_WHEN_CLOSE_CONNEXION = -1
 HYP_STATUS_NOT_ALIVE = -10
 
-
+MAX_GET_KVM_RETRIES = 3
 
 class hyp(object):
     """
@@ -58,7 +60,7 @@ def __init__(self, address, user='root', port=22, capture_events=False, try_ssh_
 
         # dictionary of domains
         # self.id = 0
-        self.domains = {}
+        self.domains = []
         if (type(port) == int) and port > 1 and port < pow(2, 16):
             self.port = port
         else:
@@ -222,7 +224,35 @@ def set_status(self, status_code):
 
             # set_hyp_status(self.hostname,status_code)
 
+    def get_kvm_mod(self):
+        for i in range(MAX_GET_KVM_RETRIES):
+            try:
+                d = exec_remote_cmd('lsmod |grep kvm',self.hostname,username=self.user,port=self.port)
+                if len(d['err']) > 0:
+                    log.error('error {} returned from command: lsmod |grep kvm'.format(d['err'].decode('utf-8')))
+                else:
+                    s = d['out'].decode('utf-8')
+                    if s.find('kvm_intel') >= 0:
+                        self.info['kvm_module'] = 'intel'
+                    elif s.find('kvm_amd') >= 0:
+                        self.info['kvm_module'] = 'amd'
+                    elif s.find('kvm') >= 0:
+                        self.info['kvm_module'] = 'bios_disabled'
+                        log.error('No kvm module kvm_amd or kvm_intel activated. You must review your BIOS')
+                        log.error('Hardware acceleration is supported, but disabled in the BIOS settings')
+                    else:
+                        self.info['kvm_module'] = False
+                        log.error('No kvm module installed. You must review if qemu-kvm is installed and CPU capabilities')
+                return True
+
+            except Exception as e:
+                log.error('Exception while executing remote command in hypervisor to list kvm modules: {}'.format(e))
+                log.error(f'Ssh launch command attempt fail: {i+1}/{MAX_GET_KVM_RETRIES}. Retry in one second.')
+            time.sleep(1)
 
+        self.info['kvm_module'] = False
+        log.error(f'remote ssh command in hypervisor {hostname} fail with {MAX_GET_KVM_RETRIES} retries')
+        return False
 
     def get_hyp_info(self):
 
@@ -281,12 +311,12 @@ def get_hyp_info(self):
         # intel virtualization => cpu feature vmx
         #   amd virtualization => cpu feature svm
         if tree.xpath('/capabilities/host/cpu/feature[@name="vmx"]'):
-            self.info['virtualization_bios_enabled'] = True
+            self.info['virtualization_capabilities'] = 'vmx'
         elif tree.xpath('/capabilities/host/cpu/feature[@name="svm"]'):
-            self.info['virtualization_bios_enabled'] = True
+            self.info['virtualization_capabilities'] = 'svm'
         else:
-            self.info['virtualization_bios_enabled'] = False
-            log.error('HYPERVISOR {} have bios vmx or svm virtualization capabilities activated??'.format(self.hostname))
+            self.info['virtualization_capabilities'] = False
+
 
     def define_and_start_paused_xml(self, xml_text):
         # todo alberto: faltan todas las excepciones, y mensajes de log,
@@ -842,13 +872,51 @@ def get_load(self):
         if raw_stats is False:
             return False
 
+        domains_with_stats = list(raw_stats['domains'].keys())
+        #broom action: domains that are started or stopped in stats that have errors in database
+        self.update_domains_started_and_stopped(domains_with_stats)
+
         self.process_hypervisor_stats(raw_stats)
 
         if len(self.stats_hyp_now) > 0:
             self.send_stats()
 
         return True
 
+    def update_domains_started_and_stopped(self,domains_with_stats):
+        if self.id_hyp_rethink is None:
+            try:
+                self.id_hyp_rethink = get_id_hyp_from_uri(hostname_to_uri(self.hostname, user=self.user, port=self.port))
+            except Exception as e:
+                log.error('error when hypervisor have not rethink id. {}'.format(e))
+                return False
+        l_all_domains = get_domains_with_status_in_list(list_status=['Started', 'Stopped', 'Failed'])
+        for d in l_all_domains:
+            if d['id'] in domains_with_stats:
+                if d['status'] == 'Started':
+                    #if status started check if has the same hypervisor
+                    if d['hyp_started'] != self.id_hyp_rethink:
+                        log.error(f"Domain {d['id']} started in hypervisor ({self.id_hyp_rethink}) but database says that is started in {d['hyp_started']} !! ")
+                        update_domain_status(status='Started',
+                                             id_domain='_admin_downloaded_tetros',
+                                             detail=f'Started in other hypervisor!! {self.id_hyp_rethink}. Updated by status thread',
+                                             hyp_id=self.id_hyp_rethink)
+                else:
+                    #if status is Stopped or Failed update, the domain is started
+                    log.info('Domain is started in {self.id_hyp_rethink} but in database was Stopped or Failed, updated by status thread')
+                    update_domain_status(status='Started',
+                                         id_domain='_admin_downloaded_tetros',
+                                         detail=f'Domain is started in {self.id_hyp_rethink} but in database was Stopped or Failed, updated by status thread',
+                                         hyp_id=self.id_hyp_rethink)
+
+            elif d['hyp_started'] == self.id_hyp_rethink:
+                #Domain is started in this hypervisor in database, but is stopped
+                if d['status'] == 'Started':
+                    update_domain_status(status='Stopped',
+                                         id_domain='_admin_downloaded_tetros',
+                                         detail=f'Domain is stopped in {self.id_hyp_rethink} but in database was Started, updated by status thread',
+                                         )
+
     def send_stats(self):
         #hypervisors
         send_stats_to_rethink = True