WEBVTT 00:00.230 --> 00:01.840 Hi, everyone, and welcome back. 00:01.850 --> 00:05.750 Now in today's video, we will be discussing about server weights. 00:05.780 --> 00:11.780 Now, server weight basically allows us to customize the request flow going between enginex that is 00:11.780 --> 00:15.200 acting as a load balancer towards the upstream backend. 00:15.200 --> 00:19.010 So let's assume that you have a load balancer here and you have two servers. 00:19.040 --> 00:22.520 Now both of these servers have a different size altogether. 00:22.520 --> 00:28.190 First server has a size of one GB Ram and second server has the capacity of four GB Ram. 00:28.190 --> 00:34.160 So definitely the first server might be able to handle a little less request when compared to that of 00:34.160 --> 00:35.240 the second server. 00:35.240 --> 00:41.900 So depending upon the size of the server, we do not really want 5,050% traffic to be equally distributed. 00:41.900 --> 00:44.020 This might not be the best strategy here. 00:44.060 --> 00:45.020 What do you want? 00:45.050 --> 00:51.590 Maybe you want 20% of the traffic to go to the smaller server and 80% of the traffic to go to the larger 00:51.590 --> 00:51.980 server. 00:51.980 --> 00:57.980 So when you have a customized environment where you need a different kind of traffic distribution, 00:57.980 --> 01:01.680 this can be achieved with the help of server weights parameter. 01:01.680 --> 01:06.780 So if you look into the sample configuration here, all you have to do is you have to specify the weight 01:06.810 --> 01:07.200 here. 01:07.200 --> 01:11.790 So for the second server with the IP of 0.4, you have a weight of two. 01:11.820 --> 01:18.510 So two requests will be sent to the IP of 0.4 and one request will be sent to the IP of 0.3. 01:18.540 --> 01:22.620 A very straightforward approach, but a very useful in production environments. 01:22.620 --> 01:25.020 So let's go ahead and get it tested. 01:25.020 --> 01:30.930 Now for today's demo, I have three servers yet again and all of these three servers have Nginx up and 01:30.930 --> 01:32.100 running perfectly well. 01:32.100 --> 01:37.590 So using the utility, let's go ahead and send ten requests here to localhost. 01:37.590 --> 01:41.400 And currently this command we are running it in a load balancer server. 01:41.400 --> 01:47.610 Now before we do that, let's go ahead and do a tale on var log Nginx access dot log. 01:47.850 --> 01:48.450 All right. 01:50.670 --> 01:54.060 Similarly, let's also do a tale here. 01:54.090 --> 01:58.170 Var log Nginx access dot log perfect. 01:58.200 --> 02:02.070 Now from the load balancer, let's go ahead and send ten requests here. 02:02.760 --> 02:09.180 Now you will see among these ten requests, five requests came to server one and five requests came 02:09.180 --> 02:09.870 to server two. 02:09.900 --> 02:14.310 So a perfect round robin based distribution of traffic is happening. 02:14.310 --> 02:15.960 So this is the perfect approach. 02:15.960 --> 02:21.810 Now let's assume that the server zero one is a smaller one and the server zero two is a larger one. 02:21.810 --> 02:25.770 So you want a little different distribution of traffic. 02:25.770 --> 02:33.420 So what you will do, you go to Nginx.conf dot D, let's modify the load balancer dot conf and for the 02:33.420 --> 02:36.720 server two, let's quickly verify the IP address again. 02:36.720 --> 02:38.700 So I'll do a control C here. 02:38.730 --> 02:40.580 Let's do a ifconfig here. 02:40.650 --> 02:48.450 The private IP address is 0.3, so for 0.3 based server I'll say weight is equal to two. 02:48.480 --> 02:50.040 I'll go ahead and save this. 02:50.160 --> 02:52.710 So let's again do a tail. 02:53.790 --> 02:54.480 All right. 02:56.580 --> 02:58.230 Same for the server one. 02:58.830 --> 03:01.910 Let's wait for new request within here. 03:01.920 --> 03:05.460 Always verify if your configuration is successful and yes it is. 03:05.460 --> 03:08.490 Let's go ahead and do a restart on Nginx. 03:08.520 --> 03:11.310 Once done, let's run the command once more. 03:11.400 --> 03:16.350 We'll send ten request and again all the ten requests are sent. 03:17.010 --> 03:23.610 Now if you look into the server one this time only three requests came to server one and all of the 03:23.610 --> 03:26.490 rest seven requests came to the server zero two. 03:26.490 --> 03:33.030 So depending upon how uneven distribution of traffic you need, you can go ahead and set the weight 03:33.030 --> 03:35.160 accordingly within your configuration file. 03:35.160 --> 03:41.340 So very simple and a very important parameter that you need to know while dealing with production traffic. 03:41.340 --> 03:43.680 So with this, we'll conclude today's video.