1 00:00:02,550 --> 00:00:09,730 So why should we implement passive node's the role of the passive notice to monitor the availability 2 00:00:09,730 --> 00:00:14,300 of the active node and the service provided by the node. 3 00:00:14,550 --> 00:00:22,690 If they are unavailable the passive node activates this requires at least two computers the active and 4 00:00:22,690 --> 00:00:24,720 passive nodes to be connected. 5 00:00:27,330 --> 00:00:32,670 They need to share a local area network which is dedicated to checking the availability of the two nodes 6 00:00:37,940 --> 00:00:43,640 best practices show that an active node availability test is of key importance for the entire fell over 7 00:00:43,640 --> 00:00:44,210 cluster 8 00:00:47,100 --> 00:00:47,670 knowingness. 9 00:00:47,690 --> 00:00:53,250 You shouldn't use a public web interface for the purpose. 10 00:00:53,500 --> 00:00:59,140 If the active node was a file server that receives and sends huge amounts of data it might not be able 11 00:00:59,140 --> 00:01:06,740 to answer an availability query in time as a consequence the service would fail over to the passive 12 00:01:06,740 --> 00:01:09,800 node which is of course something we wouldn't want 13 00:01:12,780 --> 00:01:14,890 other than the active and passive nodes. 14 00:01:15,140 --> 00:01:20,530 There are also shared resources in a service fail over cluster. 15 00:01:20,560 --> 00:01:26,610 This is a location to which an active node saves changes in its configuration. 16 00:01:26,620 --> 00:01:32,470 This implies that it's not enough to have two computers and a system that supports clustering a service 17 00:01:32,470 --> 00:01:38,200 that is protected with these measures must also support clustering. 18 00:01:38,230 --> 00:01:42,070 It has to log all changes in configuration and quorum files. 19 00:01:42,670 --> 00:01:46,820 The changes are asynchronous which means they occur only in one active node 20 00:01:49,560 --> 00:01:54,810 This explains why if there is a fail and a fail over following a change in settings the passive node 21 00:01:54,810 --> 00:01:59,740 might not acknowledge and implement the changes. 22 00:01:59,760 --> 00:02:01,920 This could lead to a system inconsistency 23 00:02:04,400 --> 00:02:12,570 the quorum which only the active node can access is capable of synchronizing both services. 24 00:02:12,610 --> 00:02:18,820 Imagine that because of a virus or detected intrusion or a simple failure we need to shut down the computer 25 00:02:18,820 --> 00:02:26,820 or stop running a service. 26 00:02:26,870 --> 00:02:30,630 What can be done to restore the services. 27 00:02:30,640 --> 00:02:34,230 There are a few options to choose from. 28 00:02:34,360 --> 00:02:42,360 The first solution which is also the safest and most recommended option is re-installation the drawback 29 00:02:42,360 --> 00:02:47,270 of re-installation is that it is unfortunately also the most time consuming solution. 30 00:02:49,610 --> 00:02:54,620 To reinstall a system we need to use disks that we've checked and are sure they have not been modified. 31 00:02:56,190 --> 00:03:02,180 We need to raise data from all hard disks of a compromised computer. 32 00:03:02,260 --> 00:03:08,630 We have to do this because we're not sure whether or not the files have been modified. 33 00:03:08,750 --> 00:03:21,310 It's far easier to delete all data and reinstall the system than find one modified file. 34 00:03:21,320 --> 00:03:27,680 The second option is recovering a computer system and installed services from backups to be able to 35 00:03:27,680 --> 00:03:31,830 recover a system we need to have made backups on a regular basis. 36 00:03:33,430 --> 00:03:38,290 The strategy for creating backup copies will be covered later. 37 00:03:38,300 --> 00:03:42,800 The files that are saved in those backups have to be trustworthy. 38 00:03:43,030 --> 00:03:47,410 We need to be fully sure that the copy has been made before an attack from which we're recovering took 39 00:03:47,410 --> 00:03:50,030 place. 40 00:03:50,230 --> 00:03:53,450 Otherwise recovering files from a backup doesn't make sense. 41 00:03:54,360 --> 00:03:58,230 This suggests that it's necessary to keep more than one backup of your infrastructure.