Questions tagged as ['load-balancing']

PM2 allows to run NodeJS apps on multiple instances, i.e., different cores, allowing load balancing using the same port.

PORT=3000 pm2 start -i NUMBER_OF_CORES(e.g 2) app.js

But I could also do load balancing in Nginx with different ports

upstream app_servers {
    server 127.0.0.1:3000;
    server 127.0.0.1:3001;
    server 127.0.0.1:3002;
    server 127.0.0.1:3002;
}

server {
    listen 80;
   ...

I want to implement HA system on azure, for that I need to move secondary IP address of active vm to standby vm as soon failover happens, how can we do this?

I am trying to list all the load balancers in a text file with tag deletion protection enabled but seems to be having issues.

$loadbalancers = aws alb describe-load-balancers --query 'LoadBalancers[*].[LoadBalancerName,DNSName,LoadBalancerArn]' --output text

foreach ($alb in $loadbalancers ) {
aws alb describe-load-balancers-attribute --loadbalancers $alb --attributes Key=deletion_protection.enabled,Value ...

If a HTTP client sends a GET request with a body that would generate a 400 Bad Request response, should the load balancer forward that request to the backend or deal with it immediately? Is there any advantage in NOT dealing with it at the load balancing layer?

Recently, an application team complained that a load balancer was returning 400 Bad Request when the application itself would return 405  ...

I have read that

Load balancers/reverse proxies usually have 2 operation modes.

In the first one, the requests from the clients are forward to one of the backends as is if they come directly from the source. Is this case the LB only redirects the request and the backend answers back directly to the client.

On the second mode, the LB answers the request and then creates a new one to the backend with  ...

I have two vms behind the Azure Load balancer working as High availability that is only one vm is providing service at a time, other vm will be in passive state. I want to keep same IP address 10.3.3.51 on both VM how can I achieve that? health probe should send probe on 10.3.3.51 to the both VMs.

I configures my TP-Links ER605 multi-wans router to load balance between my two ISPs (that is working). But since my Web/NAS server is behind my router on the 192.168.2.0 network, any computers on 192.168.0.0 and 192.168.1.0 can't access the NAS or Web server directly. The ER605 won't let use the same subnet on the the lan and the two wan.

I look at these 4 questions but no luck.

load balancing isps on  ...

I have configured a failover load balancer, so that it acts as a backup whenever my primary goes down. So I have setup Keepalived that switches the floating virtual IP address to the other machine whenever it is unable to find the service HAProxy running on other machine. The IP addresses mentioned in conf file are present on my eth1 interface.

On my primary load balancer I am getting

systemctl status ke ...

According to section 6 of RFC 3074, it seems that the process for selecting a load balancing server in ISC DHCP is as follows:

1. The client broadcasts a discover message, which either directly reaches the DHCP servers in active-active state, or which is relayed to both servers (we're ignoring the possibility of section 5.4 "HBA Intended for a Forwarder" for this example).
2. The servers assign a Serv ...

Section 2.2 of RFC 3074 defines a Service Transaction ID (STID) as:

An attribute of the individual client requests used for load-balancing.

However, section 2 of RFC 2131 defines the xid field of a DHCP packet as:

Transaction ID, a random number chosen by the client, used by the client and server to associate messages and responses between a client and a server.

These seem to match up, until  ...

I have a LB which is redirecting request in a round robin mechanism to my 4 servers configured. I have assigned a floating IP address(LOADBALANCERFLOATINGIPADDRESS) to the machine. Now I want my LB to listen on both the IP address. So I tried binding to both the IP address in listen but that does not works.

sudo nano /etc/haproxy/haproxy.cfg

#HAProxy for web servers
frontend web-frontend
  bind LOADBALAN ...

Theoretically there are:

• client
• LB
• SERVER A
• SERVER B

Question: Is there a load-balancer solution (either by software or payed in the cloud) that can redirect the "clients" to "SERVER A" or "SERVER B" based on what network port did they connect to to the LB?

If software, probably there is a solution to have firewall rules on the LB, but I am mainly searching for a payed solution, ex.: AWS, GCP or ...

I have a machine for which I have assigned a floating IP address. That machine is also my load balancer. I can access my service easily using the IP address of load balancer.

However I am unable to access it using the floating IP address which was assigned to my load balancer machine.

sudo nano /etc/haproxy/haproxy.cfg

defaults
        log     global
        mode    http
        option  httplog
   ...

I'm studying apache 2.4 and I've found this configuration as example. What I don't understand is the definition of ProxyPass, they have defined "balancer://cluster/example1/api" but in the Proxy directive there is only "balancer://cluster".

What they have done in this configuration? Can anyone explain me with some pratice example?

<VirtualHost *:10000> 
    ProxyRequests Off
    ProxyTimeout 300 ...

we have an ECS task that makes Rest API calls of two types: /path1 and /path2 in a Private subnet. We route the requests to Internet Gateway through a NAT gateway present in a public subnet.

Flow: ECS task -> NAT Gateway -> Internet Gateway.

Now, we want to route the /path2 requests to a VPC endpoint that we will create.

ECS task -> path/1 -> NAT Gateway -> Internet Gateway.

ECS task ->  ...

I have a weird scenario where I need to route certain outbound traffic from an app hosted on EC2 through different regional IP addresses.

Long story short we have 2 clients with APIs that our app communicates with to POST data. We found that their firewalls are strict at times and eventually block us mostly because we are in a different country. After a lot of back and forth, I eventually cloned  ...

I created an Internal HTTP Load balance by the code below.

I can see Internal HTTP Load balance is running with internal IP(var.my_iip)

But in VPC network > IP addresses, the "in use by" of internal IP(var.my_iip) is "none"

Why not "Forwarding rule http-ife"? and how to make it right?(to prevent someone delete this IP cuz it seem no in use)

resource "google_compute_address" "ilb-iip" {
  name   ...

This is a tough question to phrase in the limited space for the title, but I'll try my best to explain myself. I want the users of my website to be able to run the servers themselves. Those servers would be valued and rated. I want to submit those values to my load balancer so that it weights the traffic between all the servers based on how powerful they are. The load balancer should dynamically balance ...

The documentation of MetalLB states that:

In layer 2 mode, all traffic for a service IP goes to one node.

To my understanding, this is mostly due to the fact that:

one node assumes the responsibility of advertising a service to the local network.

As mentioned in the rest of the said documentation, this behavior implies a severe limitation. The traffic bandwidth is limited to what can pass through the ...

I have a seven node microk8s cluster set up on an Ubuntu server in my cellar. multipass shows me I have three IP address per VM since I created these hosts using the bridged network option. I thought having a routable IP might make it easier to expose the VMs and their services to my local network but probably I just made my platform more complex :P

I have succeeded in deploying https://ubuntu.com/tutori ...

When a browser connects to my website by entering it's FQDN the connection is encrypted using the site's SSL Certificate.

However it has come to my attention that it is possible to bypass the SSL and initiate an unencrypted connection by entering the load balancer IP Address (in which case you get an error saying it could not be encrypted).

The link below suggests that I could solve this with a redi ...

I've got an NGINX config as follows:

server {
    listen       80;
    server_name  _;
    location = / {
        proxy_pass https://server-2.alice.com/;
        proxy_redirect  off;
        proxy_ssl_server_name on;
        proxy_ssl_session_reuse off;
    }
    location /app {
        proxy_pass http://server-6.alice.com:3000$request_uri;
        proxy_redirect  off;
        proxy_ssl_server_name ...

I was reading an article about warm recoverable web server with Compute Engine and Cloud Storage.

Normal scenario:

Failover:

Documentation states that:

In an outage, you update the external HTTP(S) Load Balancing configuration and fail over to a static site in Cloud Storage.

Can the change of external HTTP(S) Load Balancing configuration occur automatically based on some health checks? For example ...

I install and configure haproxy v1.4 on solaris machine i sessufuly add two weblogic servers and i can see the uptime and downtime on mangement. my quastion is i give my weblogic url http://192.168.xx.xx:7001/pp21/loginft.jsf but the haproxy only read the url 192.168.xx.xx:7001 and it give me error. How to add full url on haproxy.cfg and access using my *:80

Error 404--Not Found From RFC 2068 Hyp ...

We are using GCP external HTTPS load balancer, architecture is shown in the diagram below. The primary use of LB is redirecting users to static error site (hosted on Cloud Storage bucket) in case CE instance is down, Traefik crashes on CE, Docker crashes on CE, etc.

We have 4 backend services defined on load balancer:

1. static-error-page backend bucket
2. blog-backend-service, gallery-backend-service ...

When I start up pound with service it does not log the error in syslog but when I start it with 'pound -d 3'

7F1384113640 thr_http start /build/pound-lNcD5I/pound-3.0/src/http.c:535 7F1384113640 start loop /build/pound-lNcD5I/pound-3.0/src/http.c:539 Listener 64.62.211.170:http: can't bind socket Listener 64.62.211.171:http: can't bind socket

Config is:

/etc/pound/pound.yaml

Minimal sample configuration

Do all response packages sent from a backend server to a client must pass the load balancer?

Actually I try to solve this Question:

I have 3 statement. 2 of 3 are true.

What is the wrong statement?

1. Do all response packages sent from a backend server to a client must pass the load balancer?

2. All incoming connections to a cluster service are received by a load balancer.

3. Load balancers can analyze traf ...

I think the subject captures my issue pretty well.

I have OpenVPN working fine behind a NLB on AWS. But, when I do the same setup (default UDP port 1194 config) on GCP, my client fails to connect.

Is there some further config (I've checked firewall rules) necessary?

Thanks.

Single Unix-based load balancer (LB) is configured with multiple network interfaces - 2x external and 1x internal. The LB sits in front of a DNS resolver. The issue I'm trying to solve is that all DNS responses come back to External Interface #1 even when the request comes through Interface #2. If the request came through Int #2, I need the response to come back through the same interface.

Update ...

We have a situation where we have multiple EC2 instances each running a VPN. Both the remote VPN server and remote subnet are run by a third-party and we have no say in the way they are setup.

We don't believe these are transferable to AWS client Lan-to-Lan VPN

The VPNs all route to the same physical subnet with the same CIDR block. There is some rate limiting for these VPNs (on the remote side) an ...