Comment 27 for bug 1413341

Hi guys, here is a little bit more info.
root@node-8:~# wget http://192.168.0.10:5000/http://192.168.0.2:5000/v2.0/tokens
--2015-02-09 22:26:43-- http://192.168.0.10:5000/http://192.168.0.2:5000/v2.0/tokens
Connecting to 192.168.0.10:5000... connected.
HTTP request sent, awaiting response... ^C
root@node-8:~# wget http://192.168.0.10:5000/v2.0/tokens
--2015-02-09 22:26:53-- http://192.168.0.10:5000/v2.0/tokens
Connecting to 192.168.0.10:5000... connected.
HTTP request sent, awaiting response... Read error (Connection timed out) in headers.
Retrying.

--2015-02-09 22:27:15-- (try: 2) http://192.168.0.10:5000/v2.0/tokens
Connecting to 192.168.0.10:5000... connected.
HTTP request sent, awaiting response... Read error (Connection timed out) in headers.
Retrying.

--2015-02-09 22:27:30-- (try: 3) http://192.168.0.10:5000/v2.0/tokens
Connecting to 192.168.0.10:5000... connected.
HTTP request sent, awaiting response... Read error (Connection timed out) in headers.
Retrying.

--2015-02-09 22:28:02-- (try: 4) http://192.168.0.10:5000/v2.0/tokens
Connecting to 192.168.0.10:5000... connected.
HTTP request sent, awaiting response... Read error (Connection timed out) in headers.
Retrying.

^C
root@node-8:~# ^C
root@node-8:~# netstat -anp |grep -c 11211
4257

I think that this is an issues with using of memcache or to be more specific with keystone to memcache connection_pool.
There are 4.5k connections to the memcache that for me means only one thing: we're not reusing them. So I will recommend to take a look in our keystone.cache.memcache_pool implementation.