# Load balancer issues
**URL:** https://forums.suse.com/t/load-balancer-issues/5275
**Category:** General
**Created:** [January 20, 2017, 4:55pm UTC](https://forums.suse.com/t/load-balancer-issues/5275 "2017-01-20T16:55:54Z")
**Posts on this page:** 4
**Page:** 1
### Author:  [@thrinay](https://forums.suse.com/u/thrinay)
#### Post date: [January 20, 2017, 4:55pm UTC](https://forums.suse.com/t/load-balancer-issues/5275/1 "2017-01-20T16:55:54Z")
When am trying to setup load balancer in rancher, I get “Failed to apply lb config on provider: error loading == reloading haproxy config with the new config changes \n[ALERT] 017/113443 (42) : parsing [/etc/haproxy/haproxy.cfg:81]: ‘rise’ has to be \> 0.\n[ALERT] 017/113443 (42) : parsing [/etc/haproxy/haproxy.cfg:82]: ‘rise’ has to be \> 0.\n[ALERT] 017/113443 (42) : Error(s) found in configuration file : /etc/haproxy/haproxy.cfg\n[WARNING] 017/113443 (42) : config : ‘option forwardfor’ ignored for proxy ‘default’ as it requires HTTP mode.\n[ALERT] 017/113443 (42) : Fatal errors found in configuration.\n: exit status 1” " which I dont have a clue of what the issue is,
Docker compose and rancher compose files of mine are as follows:
Rancher compose file:
.catalog:
name: dummy
version: "1.0.0"
description: |
Deploys a dummye project
questions:
```
- variable: docker_tag
label: Docker tag
description: Specify the docker tag to use for this deployment
default: latest
required: true
type: string
- variable: service_url
label: URL for service
description: Do not enter the TLD, AND USE TWO DASHES INSTEAD OF A DOT!!
required: true
type: string
- variable: service_type
label: Visibility
description: "Choose 'webui_external' for reachability from 'outside'"
type: enum
options:
- webui
- webui_external
required: true
```
frontend:
scale: 2
health\_check:
port: 9000
interval: 10000
unhealthy\_threshold: 2
request\_line: GET /health HTTP/1.0
health\_treshold: 3
response\_timeout: 2000
lb:
scale: 2
load\_balancer\_config:
haproxy\_config: {}
Docker-compose file:
frontend:
image: imagename
ports:
- "9000"
labels:
io.rancher.scheduler.affinity:container\_label\_soft\_ne: io.rancher.stack\_service.name=$${stack\_name}/$${service\_name}
io.rancher.container.pull\_image: always
io.rancher.container.hostname\_override: container\_name
carstool:
image: imagewhatever
ports:
- "5000"
labels:
io.rancher.scheduler.affinity:container\_label\_soft\_ne: io.rancher.stack\_service.name=$${stack\_name}/$${service\_name}
io.rancher.container.pull\_image: always
io.rancher.container.hostname\_override: container\_name
lb:
image: rancher/load-balancer-service
ports:
- "2000"
links:
- “frontend:frontend”
- "carstool:carstool"
labels:
[rancon.name](http://rancon.name): ${service\_url}
rancon.tag: ${service\_type}
io.rancher.loadbalancer.target.frontend: /=9000
io.rancher.loadbalancer.target.carstool: /api=5000
io.rancher.scheduler.affinity:container\_label\_soft\_ne: io.rancher.stack\_service.name=$${stack\_name}/$${service\_name}
io.rancher.container.pull\_image: always
io.rancher.container.hostname\_override: container\_name
---
### Author:  [@Kaelen\_Baldwin](https://forums.suse.com/u/Kaelen_Baldwin)
#### Post date: [August 7, 2017, 4:37pm UTC](https://forums.suse.com/t/load-balancer-issues/5275/2 "2017-08-07T16:37:58Z")
The main problem seems to be the rise and fall being set to 0. I’m not sure yet how to get around it, but I’m having the same issue. I’ll post here if I figure it out.
---
### Author:  [@Kaelen\_Baldwin](https://forums.suse.com/u/Kaelen_Baldwin)
#### Post date: [August 8, 2017, 4:22pm UTC](https://forums.suse.com/t/load-balancer-issues/5275/3 "2017-08-08T16:22:38Z")
problem here is in your rancher-compose.yml “health\_treshold: 3” should be “healthy\_threshold: 3”
I had the same problem, was spelling threshold as threshhold
---
### Author:  [@Kaelen\_Baldwin](https://forums.suse.com/u/Kaelen_Baldwin)
#### Post date: [August 8, 2017, 4:23pm UTC](https://forums.suse.com/t/load-balancer-issues/5275/4 "2017-08-08T16:23:17Z")
Probably you figured this out a long time ago, but for anyone else running into this issue.