invalid /etc/ctl.conf file generated when using high concurrency API requests

When executing many concurrent API requests to create targets/extents/targetextents the end result can be an invalid ctl.conf file that will not clean itself up without manual intervention.

Observe the following snippet of my config file after running many requests:

cat /etc/ctl.conf
portal-group "default" {
}
target "iqn.2011-03.lan.bitness.istgt:onegear.pvc-4a80757e-5e87-475d-826f-44fcc4719348" {
alias "onegear.pvc-4a80757e-5e87-475d-826f-44fcc4719348"
portal-group "pg1" {
tag "0x0001"
discovery-filter "portal-name"
discovery-auth-group "no-authentication"
listen "0.0.0.0:3260"
option "ha_shared" "on"
}
portal-group "pg2" {
tag "0x0002"
discovery-filter "portal-name"
discovery-auth-group "no-authentication"
listen "0.0.0.0:3261"
option "ha_shared" "on"
}
portal-group "pg1" "no-authentication"
lun "0" "onegear.pvc-4a80757e-5e87-475d-826f-44fcc4719348"
}

Notice the portal groups got written inside the target block. At best this leaves iscsi service unable to properly handle requests, at worst I think I've seen it crash the daemon before but that seems uncommon.

To semi correct the issue I've written crude scripts available here: https://github.com/democratic-csi/democratic-csi/tree/master/contrib

to run as background 'daemon' services on the server(s)