WEBVTT 00:04.350 --> 00:10.040 We have just seen this output to help our officials were spread across different broadcasts and here 00:10.050 --> 00:16.740 you see which broker is leader for particular party, for example, Broecker was added to is leader 00:16.740 --> 00:20.070 for partition with one Nola's deforming. 00:20.340 --> 00:22.170 Let's first start consumer. 00:22.170 --> 00:25.050 Let's stop this one and read all messages. 00:25.440 --> 00:26.790 Begin for that. 00:26.790 --> 00:32.820 Let's go to our Carman's list, this one, and let's read messages from topic cards from beginning. 00:33.060 --> 00:36.660 And let me also add here remaining centavos like so. 00:36.780 --> 00:39.190 But again, please notice that it is not necessary. 00:39.450 --> 00:44.880 You are still able to connect to an entire class by connecting just to single cell. 00:45.330 --> 00:52.680 That's why here in this command you see option bootstrap celebra bootstrap means initial initial server 00:52.830 --> 00:56.580 you connect to and afterwards you connect to the entire class. 00:57.150 --> 00:58.860 OK, so let me copy this. 00:58.860 --> 01:00.320 Come and go here. 01:00.510 --> 01:03.990 Clear terminal and based command here and execute it. 01:04.320 --> 01:10.830 And I'll get all five messages that were produced by a producer here and notice again that all those 01:10.830 --> 01:15.360 messages were received not in the same order as we will produce them. 01:15.600 --> 01:21.330 And I have explained it to you in a previous example or that we won't be the same if you read from multiple 01:21.330 --> 01:22.050 partitions. 01:22.290 --> 01:26.540 And in our case, our topic cards has five different partitions. 01:27.120 --> 01:30.150 But again, notice that we have received all messages. 01:30.330 --> 01:33.320 It means that all brokers are up and running. 01:33.870 --> 01:40.620 Now, let's stop this consumer like soul and let's simulate failure of one of the brokers. 01:41.400 --> 01:47.910 Afterwards, we will start consumer again and see whether we will be able to receive all messages or 01:47.910 --> 01:48.250 not. 01:48.810 --> 01:55.800 So let's go, for example, to this broker and this broker, his idea to and let's simply stop it. 01:56.130 --> 01:57.060 Control, see. 02:02.430 --> 02:08.940 Broker will stopped and there are only two brokers with ÉDITH zero and one in our class, so we can 02:08.940 --> 02:11.410 verify that quickly using this comment. 02:11.430 --> 02:13.520 Let me scroll up this one. 02:13.740 --> 02:18.520 So at least the ideas of brokers and the people who in this comment based here. 02:19.780 --> 02:26.620 And I'll see Arae only of two, Idiz zero and one, and that's fully correct because broker with the 02:26.620 --> 02:27.740 two is down. 02:28.440 --> 02:33.910 OK, let's now go to Consumer Step and try to read messages again from beginning. 02:34.240 --> 02:36.390 So same command and. 02:38.810 --> 02:44.960 And I got following up, if I screwed up, I will see that three messages were received Dassler, Honda 02:44.980 --> 02:45.080 and. 02:45.980 --> 02:53.800 And also I see a bunch of warnings that indicate the two partitions have leader brokers without imagining 02:53.870 --> 02:57.920 Lisner, including cars four and cars one. 02:58.400 --> 03:04.670 And this warning tells us that there are two partitions that actually are not available and we are not 03:04.670 --> 03:07.220 able to read messages from those partitions. 03:07.880 --> 03:14.000 And after starp of this consumer, you'll see that in total recall protest, only three messages instead 03:14.000 --> 03:14.560 of five. 03:14.960 --> 03:17.570 It means the two messages were simply lost. 03:18.170 --> 03:20.330 And the reason for that is pretty, pretty simple. 03:20.690 --> 03:28.390 There were three bloggers and each of them had part of old messages and there was no indication of farter. 03:28.400 --> 03:34.610 And that means that two remaining messages that we haven't received here by this consumer, we are simply 03:34.610 --> 03:35.150 lost. 03:35.330 --> 03:36.820 When Broecker went down. 03:37.280 --> 03:38.470 Let's not to the following. 03:38.750 --> 03:41.870 Let's try to read the details about Karl's topic. 03:41.870 --> 03:45.830 Now, here you see previous output when all brokers were alive. 03:46.040 --> 03:47.750 So let's repeat the same comment. 03:47.900 --> 03:53.990 If you don't have to step open, please go to this file and copy a command that will read the details 03:53.990 --> 03:56.990 about specific topic, this one and executed. 03:57.110 --> 03:58.160 So after. 03:59.740 --> 04:06.670 And now I see here following big mortician's zero, two and three are still available, and here you 04:06.670 --> 04:10.510 see correspondant leader ads for them here, here and here. 04:10.960 --> 04:17.350 But partitions, one and four are not available and leader is marked here as none. 04:17.830 --> 04:22.150 And that means that those partitions are technically not available right now. 04:22.660 --> 04:27.530 Now, let's suppose that after a short period of outage, Broecker two will come up again. 04:27.670 --> 04:28.960 So let's start over. 04:29.170 --> 04:32.160 Go to Carmine's list and copied this comment. 04:32.170 --> 04:38.930 It was the start of Breuker with decided to let me get you to go to this step where we have started 04:39.070 --> 04:42.760 Broecker with a little let me clear it and paste command. 04:44.780 --> 04:45.470 Starting. 04:46.450 --> 04:52.990 And Broecker was successfully launched again, let's go back to the STEIB, where we have details about 04:52.990 --> 04:55.660 a specific topic and let's read them again. 04:55.900 --> 04:57.730 So details about Tropicana's. 05:00.310 --> 05:08.380 And now I see again that leader with Ideato is active for partizan one and for and situation is back 05:08.380 --> 05:09.000 to normal. 05:09.280 --> 05:13.990 Let's go to Consumer back and let's launch it again and read messages from beginning. 05:15.630 --> 05:21.720 And now we have again read the old file that's just like that in our current topic, and we don't receive 05:21.720 --> 05:22.860 any warnings. 05:23.460 --> 05:29.310 That's what happens when some of the brokers goes down and afterwards, after some period of time, 05:29.580 --> 05:30.870 goes up again. 05:31.230 --> 05:38.170 And notice that in our case, we have actually lost some messages because they were stored without any 05:38.240 --> 05:39.960 replication in our case. 05:39.960 --> 05:44.340 And this example, replication of fact or for this particular topic is one. 05:44.520 --> 05:48.210 And that means that messages are not obligated to other brokers. 05:48.510 --> 05:49.090 Wonderful. 05:49.290 --> 05:50.500 That's all for this example. 05:50.520 --> 05:53.290 And next, let's quickly summarize what we did here. 05:53.460 --> 05:54.120 See you next. 05:54.130 --> 05:54.590 Bye bye.