Alpine-tacstack localhost vs 127.0.0.1

[cbautista1002]

In order to get PHPMyAdmin to work I had to change its config file /etc/phpmyadmin/config.inc.php. I had to change the host value, $cfg['Servers'][$i]['host'], from localhost to 127.0.0.1.

I checked /etc/hosts and it contains 127.0.0.1 localhost.

@jumanjiman Is there a reason why 127.0.0.1 works but localhost does not? The old dockerfile had RUN echo "NETWORKING=yes" > /etc/sysconfig/network in it (6ea6b8b#diff-3254677a7917c6c01f55212f86c57fbfL23) and the new one does not. Is this needed?

CC: @gganeshan

[jumanjiman]

the echo for /etc/sysconfig/network was a redhat anachronism for their init scripts.
we don't need it since we use supervisord to start the daemons.

as far as localhost vs 127.0.0.1, it had not occurred to me to think about that aspect.
here's what happens:

1. "myapp" wants to reach "localhost".
2. the stub resolver checks /etc/nsswitch.conf to see who to use as the real resolver. options typically include files (aka /etc/hosts), dns, ldap, etc.
3. the stub resolver then does its thing against the specified service.

our alpine config does not have an /etc/nsswitch.conf, therefore we fall through to the default dns.
this is a safe, secure choice in the typical case.

example on my ec2 instance(s):

user@wormhole:~$ grep -i docker .bashrc
export DOCKER_HOST='tcp://192.168.254.162:2375'
alias dosh="docker run --rm -it alpine:latest sh"

user@wormhole:~$ dosh
# no nsswitch config...
/ # stat /etc/nsswitch.conf
stat: can't stat '/etc/nsswitch.conf': No such file or directory

# yet ping still works!
/ # ping -c1 localhost
PING localhost (127.0.0.1): 56 data bytes
64 bytes from 127.0.0.1: seq=0 ttl=64 time=0.045 ms

--- localhost ping statistics ---
1 packets transmitted, 1 packets received, 0% packet loss
round-trip min/avg/max = 0.045/0.045/0.045 ms

# install and use dig to see what dns gives us...
# NOTE in particular the "authority" that gives us a valid answer
/ # apk add --update bind-tools &> /dev/null
/ # dig localhost

; <<>> DiG 9.10.1-P1 <<>> localhost
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 35876
;; flags: qr aa rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 1

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 4096
;; QUESTION SECTION:
;localhost.                     IN      A

;; ANSWER SECTION:
localhost.              86400   IN      A       127.0.0.1

;; Query time: 0 msec
;; SERVER: 192.168.254.2#53(192.168.254.2)
;; WHEN: Wed Mar 25 14:43:47 UTC 2015
;; MSG SIZE  rcvd: 54

so, if ISE dns does not give you a valid answer for localhost (which RFC says DNS MUST do), then queries to localhost fail.

[cbautista1002]

So ISE's DNS failed to provide an answer right? Is this something that can be taken care of? Shall I create an issue in the issues repo to have this fixed?

[jumanjiman]

So ISE's DNS failed to provide an answer right?

you can check from a rhel node by running:

# both of these should return 127.0.0.1 with an ise authority
dig localhost @<ip-of-nameserver-listed-in-/etc/resolv.conf>
dig localhost.localdomain @<ip-of-nameserver-listed-in-/etc/resolv.conf>

Shall I create an issue in the issues repo to have this fixed?

if your dig shows it to be a problem, then yes

[cbautista1002]

127.0.0.1 is return for localhost @ip, but not for localhost.localdomain @ ip:

redacted

[cbautista1002]

I wanted to use supervisor's http control but in order to get that to work I had to change /etc/supervisord.conf:

[inet_http_server]
port = 127.0.0.1:9001

changed to

[inet_http_server]
port = *:9001