studio sign in button doesn't work after ec2 stop and then start

[ODS_steel]

Hi,

I am running ironwood.master (ironwood.2) on an ubuntu 16.04 aws ec2. 

EC2s can be stopped and started (I don't mean rebooted i.e. sudo reboot). After doing this to my instance I found that the sign in button to studio would consistently stop working.

Upon starting up the previously stopped instance - the console shows that one is assigned a new ipv4 public address [the private ip address asigned from the creation of the instance is unchanged after the start]

Here is an excerpt from /edx/var/log/rabbitmq/edx.cms.core.default.log  (the file is basically filled with these lines]

date='Wed Jul  3 09:34:02 UTC 2019' vhost='/' queue='edx.cms.core.default' length=0

date='Wed Jul  3 09:35:03 UTC 2019' vhost='/' queue='edx.cms.core.default' length=0

date='Wed Jul  3 09:36:02 UTC 2019' vhost='/' queue='edx.cms.core.default' length=0

date='Wed Jul  3 09:37:03 UTC 2019' vhost='/' queue='edx.cms.core.default' length=0

date='Wed Jul  3 09:38:02 UTC 2019' vhost='/' queue='edx.cms.core.default' length=0

date='Wed Jul  3 09:39:03 UTC 2019' vhost='/' queue='edx.cms.core.default' length=0

date='Wed Jul  3 09:40:02 UTC 2019' vhost='/' queue='edx.cms.core.default' length=0

date='Wed Jul  3 09:41:03 UTC 2019' vhost='/' queue='edx.cms.core.default' length=0

date='Wed Jul  3 09:42:02 UTC 2019' vhost='/' queue='edx.cms.core.default' length=0

date='Wed Jul  3 09:43:03 UTC 2019' vhost='/' queue='edx.cms.core.default' length=0

date='Wed Jul  3 09:44:02 UTC 2019' vhost='/' queue='edx.cms.core.default' length=0

date='Wed Jul  3 09:45:03 UTC 2019' vhost='/' queue='edx.cms.core.default' length=0

date='Wed Jul  3 09:46:02 UTC 2019' vhost='/' queue='edx.cms.core.default' length=0

date='Wed Jul  3 09:47:03 UTC 2019' vhost='/' queue='edx.cms.core.default' length=0

date='Wed Jul  3 09:48:02 UTC 2019' vhost='/' queue='edx.cms.core.default' length=0

date='Wed Jul  3 09:49:03 UTC 2019' vhost='/' queue='edx.cms.core.default' length=0

Here is an excerpt from /edx/var/log/cms/edx.log (the file is basically filled with these lines] [please note the numbers associated with the ip have been replaced with asterisks

Jul  3 09:31:25 ip-***-**-**-*** [service_variant=cms][enterprise.utils][env:sandbox] WARNING [ip-***-**-**-***  12901] [utils.py:50] - Could not import Registry from third_party_auth.prov$

Jul  3 09:31:25 ip-***-**-**-*** [service_variant=cms][enterprise.utils][env:sandbox] WARNING [ip-***-**-**-***  12901] [utils.py:51] - cannot import name EnterpriseCustomerUser

Jul  3 09:31:25 ip-***-**-**-*** [service_variant=cms][enterprise.utils][env:sandbox] WARNING [ip-***-**-**-***  12802] [utils.py:50] - Could not import Registry from third_party_auth.prov$

Jul  3 09:31:25 ip-***-**-**-*** [service_variant=cms][enterprise.utils][env:sandbox] WARNING [ip-***-**-**-***  12802] [utils.py:51] - cannot import name EnterpriseCustomerUser

Jul  3 09:31:27 ip-***-**-**-*** [service_variant=cms][enterprise.utils][env:sandbox] WARNING [ip-***-**-**-***  12883] [utils.py:50] - Could not import Registry from third_party_auth.prov$

Jul  3 09:31:27 ip-***-**-**-*** [service_variant=cms][enterprise.utils][env:sandbox] WARNING [ip-***-**-**-***  12883] [utils.py:51] - cannot import name EnterpriseCustomerUser

I have been getting around this (while in development stage) by just logging in via the django admin and then clicking the "view site" link which is in the right hand corner once successfully signed into django admin. 

While this is a way to get round during development, it is not a long-term solution. Any ideas on how I can fix this?

Thanks