WEBVTT 00:00.720 --> 00:02.610 -: In the last section, we started taking a look 00:02.610 --> 00:05.340 at how we can get Docker Compose to automatically maintain 00:05.340 --> 00:07.140 and restart our containers. 00:07.140 --> 00:08.610 We added in a little bit of code 00:08.610 --> 00:10.680 to our node server to make sure 00:10.680 --> 00:12.420 that anytime we visited our route 00:12.420 --> 00:14.580 the server would automatically exit. 00:14.580 --> 00:15.900 When we visited that route 00:15.900 --> 00:17.942 we were then able to flip on over 00:17.942 --> 00:19.770 to our terminal and see that our container stopped. 00:19.770 --> 00:22.020 We were also able to run Docker ps 00:22.020 --> 00:23.280 and verify that, yep 00:23.280 --> 00:25.410 we only have one container now running. 00:25.410 --> 00:26.670 So in this section we're gonna figure out 00:26.670 --> 00:28.320 exactly how we can get Docker Compose 00:28.320 --> 00:31.743 to automatically restart that crashed or stopped container. 00:32.580 --> 00:34.620 Now the first thing I wanna tell you about 00:35.689 --> 00:37.230 is that little line of code that we added inside 00:37.230 --> 00:39.810 of our index.js file inside of here. 00:39.810 --> 00:41.910 We added in process.exit 00:41.910 --> 00:44.910 and passed in the number zero right there. 00:44.910 --> 00:46.410 The zero that we're adding in right there 00:46.410 --> 00:49.650 is what we refer to as a exit status code. 00:49.650 --> 00:51.930 By adding in a zero, we are indicating 00:51.930 --> 00:54.180 that we just exited our node server 00:54.180 --> 00:56.670 and we exited because we wanted to. 00:56.670 --> 00:57.690 Everything is okay. 00:57.690 --> 00:59.220 We didn't run into any errors. 00:59.220 --> 01:00.750 We stopped that running process 01:00.750 --> 01:02.700 because we meant to. 01:02.700 --> 01:06.390 if we added in a status code of anything besides zero 01:06.390 --> 01:10.367 so 1, 2, 3, 300, 400, 500, 01:10.367 --> 01:12.660 5,000, any number besides zero 01:12.660 --> 01:14.820 that means that our running process exited 01:14.820 --> 01:18.300 because an error occurred or something went wrong. 01:18.300 --> 01:21.150 Now that's a very important little tidbit to be aware of 01:21.150 --> 01:24.330 because it's going to affect exactly how Docker decides 01:24.330 --> 01:26.460 whether or not to restart our containers. 01:26.460 --> 01:29.400 So just keep that in your mind for just a moment. 01:29.400 --> 01:31.320 Okay, so here's how we're going to get Docker Compose 01:31.320 --> 01:33.570 to automatically restart a container 01:33.570 --> 01:36.870 we're going to specify something called a restart policy 01:36.870 --> 01:39.280 inside of our Docker Compose file 01:40.663 --> 01:42.374 there are four different restart policies. 01:42.374 --> 01:43.207 So we have access to. 01:43.207 --> 01:44.040 By default 01:44.040 --> 01:46.530 we have the no restart policy assigned 01:46.530 --> 01:48.030 to all of our containers. 01:48.030 --> 01:50.760 The no restart policy means if this thing crashes 01:50.760 --> 01:54.900 for any reason, do not attempt to restart the container. 01:54.900 --> 01:57.030 We also get the always restart policy 01:57.030 --> 01:58.470 which as you might guess means 01:58.470 --> 02:01.590 if the container stops for absolutely any reason whatsoever 02:01.590 --> 02:04.080 automatically attempt to restart it. 02:04.080 --> 02:04.913 We get two others. 02:04.913 --> 02:06.210 So we'll talk about in just a moment, 02:06.210 --> 02:07.050 but before we do 02:07.050 --> 02:09.660 let's try adding in the always restart policy 02:09.660 --> 02:10.950 to our node server 02:10.950 --> 02:12.360 and then try visiting it again 02:12.360 --> 02:14.130 and trying to get it to crash. 02:14.130 --> 02:15.750 So let's get to it. 02:15.750 --> 02:17.940 I'm gonna flip back on my code editor. 02:17.940 --> 02:20.910 I'm gonna open up my Docker Compose file 02:20.910 --> 02:23.070 and then I'm going to add a new line right 02:23.070 --> 02:26.070 after our node app specification right here. 02:26.070 --> 02:28.593 I'm gonna say restart always. 02:29.610 --> 02:30.443 So as you might guess, 02:30.443 --> 02:33.540 this specifies the always restart policy 02:33.540 --> 02:36.030 for our node app container. 02:36.030 --> 02:38.520 Notice how we specified a restart policy 02:38.520 --> 02:40.350 for one specific service 02:40.350 --> 02:43.290 or one specific container. 02:43.290 --> 02:45.870 If we wanted to add a restart policy to the writer's server 02:45.870 --> 02:49.440 we would have to add it onto that one specifically as well. 02:49.440 --> 02:52.440 But right now we'll go with just node app. 02:52.440 --> 02:53.670 Now I'm gonna save this file 02:53.670 --> 02:54.660 and we're gonna try bringing 02:54.660 --> 02:57.390 up our Docker Compose containers again 02:57.390 --> 03:00.843 and seeing what happens when our node application crashes. 03:01.980 --> 03:04.350 Okay, so I'm gonna flip back on my terminal. 03:04.350 --> 03:06.647 I'm going to hit control C to stop that 03:06.647 --> 03:08.650 Docker Compose process 03:09.958 --> 03:10.800 and then I will start them back up with 03:10.800 --> 03:12.183 Docker Compose up. 03:15.600 --> 03:17.613 We'll now flip open our browser again. 03:18.450 --> 03:19.980 I'll open a new tab. 03:19.980 --> 03:21.720 I do encourage you to make a new tab here 03:21.720 --> 03:24.720 as opposed to just refreshing the existing one you have. 03:24.720 --> 03:28.020 Sometimes Chrome does not quite refresh the server 03:28.020 --> 03:31.110 as you would expect if you reuse your tab. 03:31.110 --> 03:33.933 So I'm going to visit local host :4,001. 03:35.976 --> 03:37.470 Again, I get the error message here 03:38.666 --> 03:40.727 which is okay if I flip backward my terminal 03:40.727 --> 03:41.560 after a second and two, you'll notice 03:41.560 --> 03:44.460 that we have gotten that exit message right there. 03:44.460 --> 03:45.870 But then very quickly after that 03:45.870 --> 03:48.270 we got this new color here that essentially means 03:48.270 --> 03:51.360 that we just restarted that running container. 03:51.360 --> 03:52.590 So we've restarted the container 03:52.590 --> 03:55.590 and we again have a copy of node js listing 03:55.590 --> 03:58.620 on port 81 inside the container. 03:58.620 --> 03:59.970 One thing that's kind of interesting here is 03:59.970 --> 04:02.670 so you'll notice that we are seeing multiple listening on 04:02.670 --> 04:05.400 Port 81 messages here. 04:05.400 --> 04:06.420 That's totally fine. 04:06.420 --> 04:08.910 Technically, when you and I are stopping our container 04:08.910 --> 04:10.830 with that process.exit line 04:10.830 --> 04:14.520 we're not actually deleting the container or stopping it 04:14.520 --> 04:16.920 or anything like or killing it or anything like that. 04:16.920 --> 04:19.110 Technically the container just stopping 04:19.110 --> 04:22.320 and when Docker Compose decides to restart the container 04:22.320 --> 04:25.830 it's reusing the one that we had previously created. 04:25.830 --> 04:27.790 And so when we reuse that container 04:29.208 --> 04:30.960 we attach to the standard out log 04:30.960 --> 04:33.870 which has all the previous messages appended on it as well. 04:33.870 --> 04:35.550 And that's why we're seeing multiple listing 04:35.550 --> 04:37.263 on port 81 right here. 04:38.670 --> 04:41.430 All right, so back over to our diagrams. 04:41.430 --> 04:43.830 So as you might guess, the always restart policy says, 04:43.830 --> 04:46.230 'Hey if this container stops for any reason 04:46.230 --> 04:47.730 just go ahead and attempt to restart it.' 04:47.730 --> 04:51.510 And we just saw that in action when we crashed our server. 04:51.510 --> 04:54.600 Now we also get access to the on failure restart policy. 04:54.600 --> 04:55.860 As you might guess, 04:55.860 --> 04:57.030 the on failure policy 04:57.030 --> 05:00.180 is only going to attempt to restart the running container 05:00.180 --> 05:02.160 if we get a error code. 05:02.160 --> 05:03.390 So a status code 05:03.390 --> 05:07.230 from that process.exit line other than on zero. 05:07.230 --> 05:08.063 Let's try testing that 05:08.063 --> 05:10.190 out right now and just seeing what happens. 05:11.558 --> 05:14.460 Now, I'm going to leave process.exit inside 05:14.460 --> 05:16.830 of index.js at zero right here. 05:16.830 --> 05:19.110 Okay, so we're still going to have an exit code of zero 05:19.110 --> 05:22.380 which means we exited because we wanted to. 05:22.380 --> 05:24.300 But inside of my Docker Compose file 05:24.300 --> 05:26.470 I'm gonna change the restart policy to be 05:28.664 --> 05:29.497 on - failure 05:30.480 --> 05:31.313 like so. 05:32.370 --> 05:34.380 Now I'll go back over to my terminal, 05:34.380 --> 05:36.963 I'll stop that running process with control C. 05:38.571 --> 05:39.990 I'll then start my Docker Compose group back up 05:39.990 --> 05:41.463 with Docker Compose up. 05:43.680 --> 05:46.380 Now if I open up my browser again 05:46.380 --> 05:50.373 I'll again create a new tab and visit local host 4001. 05:51.960 --> 05:53.523 I don't get any feedback here. 05:54.635 --> 05:55.468 If I go to my terminal, 05:55.468 --> 05:58.080 I'll see that I have an exit with status code Zero 05:58.080 --> 06:00.300 and now in this case we say that, 06:00.300 --> 06:02.160 'Hey we exited because we wanted to' 06:02.160 --> 06:03.600 No error whatsoever 06:03.600 --> 06:06.210 but we specified a restart policy that says, 06:06.210 --> 06:08.850 'Only attempt to restart it if we had a failure.' 06:08.850 --> 06:09.683 So in this case 06:09.683 --> 06:12.090 we're never going to see the container restart. 06:12.090 --> 06:14.310 We would have to change our exit code 06:14.310 --> 06:17.100 in order to see it restarted when using the 06:17.100 --> 06:18.513 on failure policy. 06:19.410 --> 06:21.270 Now, if you want to test that out yourself 06:21.270 --> 06:22.103 feel free to do so. 06:22.103 --> 06:24.090 You could change the exit of zero right here 06:24.090 --> 06:27.810 to one or 10 or 100 06:27.810 --> 06:30.480 or essentially any non-zero number. 06:30.480 --> 06:32.180 Just remember that if you do that, 06:33.163 --> 06:34.913 you will have to rebuild your image 06:35.979 --> 06:39.930 by using Docker Compose up dash dash built like so. 06:39.930 --> 06:41.430 So again, if you wanna try it yourself 06:41.430 --> 06:43.330 just don't forget the dash dash built. 06:44.580 --> 06:46.530 Now, the last policy that I wanna tell you 06:47.593 --> 06:49.830 about is the unless stopped policy. 06:49.830 --> 06:51.300 So unless stopped means 06:51.300 --> 06:53.380 always attempt to restart this container 06:54.327 --> 06:56.010 unless you or I at the command line 06:56.010 --> 06:58.050 forcibly tell that container to stop 06:58.050 --> 07:00.120 by running Docker stop. 07:00.120 --> 07:02.182 So unless stopped right here 07:02.182 --> 07:03.030 basically means just always restart it. 07:03.030 --> 07:04.980 But if you and I really want it to stop, 07:04.980 --> 07:06.213 then it will stop. 07:07.590 --> 07:10.530 Now, one quick note here about the 'no' restart policy 07:10.530 --> 07:11.550 you'll notice it in this case 07:11.550 --> 07:12.810 we put quotes around, 07:12.810 --> 07:15.120 'No' that's actually on purpose. 07:15.120 --> 07:16.740 So with all the other policies 07:16.740 --> 07:18.960 like always on failure and unless stopped, 07:18.960 --> 07:21.210 you can just type in the raw policy name 07:21.210 --> 07:22.950 into your Docker Compose file. 07:22.950 --> 07:25.230 So on failure like that right there. 07:25.230 --> 07:28.470 But if you're going to make use of the 'no' restart policy 07:28.470 --> 07:30.450 you specifically have to put it in quotes 07:30.450 --> 07:32.670 either double or single. 07:32.670 --> 07:35.220 The reason for that is that in a YAML file, 07:35.220 --> 07:36.240 the value no 07:36.240 --> 07:37.350 and you can see when I type it in 07:37.350 --> 07:39.360 it appears as orange right here. 07:39.360 --> 07:40.230 So in a YAML file, 07:40.230 --> 07:43.110 the value no gets interpreted as false. 07:43.110 --> 07:45.090 So a false restart policy is 07:45.090 --> 07:47.520 different than a string of no. 07:47.520 --> 07:51.472 So that's why 'no' specifically needs to be added in quotes 07:51.472 --> 07:53.160 so that it understands this is not a Boolean false 07:53.160 --> 07:55.103 it's specifically the string, 'no.' 07:56.490 --> 07:59.280 Okay, so that's pretty much it on restart policies. 07:59.280 --> 08:02.033 Now, the last thing you might be a little bit curious about 08:02.033 --> 08:03.480 is why would we ever want to use always 08:03.480 --> 08:05.460 versus on failure? 08:05.460 --> 08:06.330 Well, it really comes down 08:06.330 --> 08:08.610 to the purpose of your Docker container. 08:08.610 --> 08:09.570 In some cases 08:09.570 --> 08:12.510 you might have a container that you always, always 08:12.510 --> 08:14.970 always want to make sure is running. 08:14.970 --> 08:17.370 A good example of this would be a web server. 08:17.370 --> 08:19.710 If you are running some public web application 08:19.710 --> 08:21.540 chances are a hundred percent 08:21.540 --> 08:24.390 of the time you want that server to be available. 08:24.390 --> 08:26.040 And so if you are running some type 08:26.040 --> 08:27.840 of web application, well 08:27.840 --> 08:30.540 I would kind of expect you to use the always restart policy 08:30.540 --> 08:31.952 so that you are 08:31.952 --> 08:34.560 at least always attempting to keep your server running. 08:34.560 --> 08:35.490 On the other hand 08:35.490 --> 08:39.030 if you are running some type of worker process 08:39.030 --> 08:42.042 like some container that is meant to do some amount 08:42.042 --> 08:44.310 of processing on some file and then naturally exit 08:44.310 --> 08:46.210 that would probably be a good use case 08:47.468 --> 08:48.570 for the on failure policy 08:48.570 --> 08:49.920 because that worker container might eventually 08:49.920 --> 08:51.003 finish its job, 08:51.899 --> 08:52.732 it might finish processing a file 08:52.732 --> 08:54.840 and when it's completed processing that file 08:54.840 --> 08:56.730 Well, you probably don't wanna start it back 08:56.730 --> 08:58.170 up because it finished its job 08:58.170 --> 09:00.000 and you should probably just let it die 09:00.000 --> 09:02.940 and let it close out and not get restarted. 09:02.940 --> 09:05.550 And so if you have a worker process of some sort 09:05.550 --> 09:07.680 well that's probably where I would expect you to use the 09:07.680 --> 09:09.600 on failure policy. 09:09.600 --> 09:10.650 Okay, so that's pretty much it. 09:10.650 --> 09:12.900 That's how we get Docker Compose to automatically 09:12.900 --> 09:14.220 maintain our containers. 09:14.220 --> 09:15.630 Now, there's one last little note 09:15.630 --> 09:17.220 that I wanna share around Docker Compose. 09:17.220 --> 09:18.630 So let's take a quick break right here 09:18.630 --> 09:20.403 and continue in the next section.