Apigee 9 Node setup on Azure| Message processor goes down after virtual host update

[#1]

Hi,

Scenario:

We have setup Apigee on a VM hosted at Azure with a 9 Node topology configuration.Since, the message router sits on two nodes in this configuration, the config file is populated with these two IPs (private) and corresponding ports while giving the virtual host details. Also, "hostname -i" returns the private IPs of these routers, which is how it should be.

Issue:

1. When we make the changes in the config file to update organization, environment & virtual host alias and hit the message router IP with the updated port, the proxy was not reachable.

Config file:

IP1=********

MSIP="$IP1"

#Admin details ADMIN_EMAIL="*****"

APIGEE_ADMINPQ=*******

#Organization details

ORG_NAME=*****

ORG_ADMIN=*******

ENV_NAME=qa

VHOST_PORT=9001

VHOST_NAME=test

VHOST_ALIAS="11.1.1.5:9001 11.1.1.6:9001"

Therefore, we tried updating the virtual host alias through the following CURL command and the router IP was reachable with the new port number.

http://11.1.1.1:8000/v1/organizations/b2b/environments/qa/virtualhosts/test \

-d '<VirtualHost name="test"> \

<HostAliases> \

<HostAlias>11.1.1.5:9001</HostAlias>

<HostAlias>11.1.1.6:9001</HostAlias> \

</HostAliases> \

<Interfaces/> \

<Port>9001</Port> \

</VirtualHost>

So, in crux the CURL command updated the details, whereas the manual config file update didn't work.

Post this,

2. Suppose, the virtual host details are again updated, the latest change in the virtual host alias works fine but the previous port does not respond to any proxy calls. Also, the Message processor goes down after this update.

3. We have to configure a load balancer on these IPs. Require urgent support on that as well.

Here are the log files

logs.txt

[#2]

@testuser, How many Mps do you have?
As per logs, the MP on 52.172.214.81 did not start. Will need complete logs to understand why it did not start.

BTW, you can configure the router IPs on the load balancer and just have a hostname in the Hostalias in the virtualhost which should be in the HOST header that comes from the LB.

More info here http://docs.apigee.com/api-services/content/virtual-hosts#creatingavirtualhost

[paulmibus]

This message may indicate the cause:

com.apigee.errors.http.server.ServiceFailureException: Failed to start server on "/52.172.214.81:8998"

That IP address is a public address; most cloud services provision private (RFC 1918) addresses to instances and use public addresses to NAT traffic to the private addresses. Can you check the message processor to confirm that 52.172.214.81 is assigned to an interface using this command?

> ip addr show

[#4]

@Paul Mibus We are facing issues while accessing the VM, which is why we cannot use these commands as well.

[#5]

@Maruti Chand We have two MPs.

Sharing the logs again. logs-apigee.txt

[#6]

@testuser Pls, check the answer from @Paul Mibus.

You should check that first and share the complete setup-root.log and system.log.

[#7]

@Maruti Chand

Sharing the logs. The message processor goes down after a virtual host update.

system.txt

edge-message-processor.txt

[#8 · Accepted solution]

Thanks a lot for the support folks. Looks like there was an IP conflict that was causing the issue. Works fine now. Thanks for keeping me posted for help here.