Occasionally you may run into an issue where the routes between the routers and nats are not in a consistent state. To detect this issue log onto a server with connectivity to your BOSH director and copy the following Bash script onto it:
#!/usr/bin/env bash
DEPLOYMENT=$1
RUSER=$2
RPASSWORD=$3
if [[ -z ${DEPLOYMENT} ]]; then
echo "No deployment specified."
exit 1
fi
if [[ -z ${RUSER} ]]; then
echo "No router user specified"
exit 1
fi
if [[ -z ${RPASSWORD} ]]; then
echo "No router user password specified"
exit 1
fi
if ! [ -x "$(command -v jq)" ]; then
echo 'jq is not installed and is requried - please install' >&2
exit 2
fi
for ip in $(bosh vms $DEPLOYMENT| grep router_z | grep -oE "[0-9]+\..*[0-9]+"); do
echo Router IP Address: $ip
temp_length=$(curl -s http://$RUSER:[email protected]$ip:8080/routes | jq length)
echo Router Routes length: $temp_length
temp_length=
done
To use this script you will need to retrieve the router username and password from your deployment manifest and call the script with the following parameters:
script <deployment name> <router.status.user> <router.status.password>
This will generate output similar to:
Router IP Address: 10.8.8.100
Router Routes length: 12121
Router IP Address: 10.8.8.101
Router Routes length: 12121
Router IP Address: 10.8.8.102
Router Routes length: 4336
If there are inconsistencies you likely just need to monit restart
on all of the nats
servers.