Ad hoc commands are not returning to the prompt?

Archives Ansible Developer
1

Hi everyone,

Just started out using Ansible for managing our various hosts / docker containers etc… The only problem i have whilst testing this: all my ad-hoc commands never finish :slight_smile:

just a simple ‘ansible office -vvvvv -a “/usr/bin/uptime”’
ESTABLISH CONNECTION FOR USER: root on PORT 22 TO host1
ESTABLISH CONNECTION FOR USER: root on PORT 22 TO host2
ESTABLISH CONNECTION FOR USER: root on PORT 22 TO host3
ESTABLISH CONNECTION FOR USER: root on PORT 22 TO host4
EXEC /bin/sh -c ‘mkdir -p $HOME/.ansible/tmp/ansible-1399929485.48-193948231594553 && echo $HOME/.ansible/tmp/ansible-1399929485.48-193948231594553’
EXEC /bin/sh -c ‘mkdir -p $HOME/.ansible/tmp/ansible-1399929485.52-91011211544401 && echo $HOME/.ansible/tmp/ansible-1399929485.52-91011211544401’
REMOTE_MODULE command /usr/bin/uptime
REMOTE_MODULE command /usr/bin/uptime
PUT /tmp/tmpqfWmwp TO /root/.ansible/tmp/ansible-1399929485.48-193948231594553/command
EXEC /bin/sh -c ‘mkdir -p $HOME/.ansible/tmp/ansible-1399929485.49-54151257537199 && echo $HOME/.ansible/tmp/ansible-1399929485.49-54151257537199’
PUT /tmp/tmplYjZje TO /root/.ansible/tmp/ansible-1399929485.52-91011211544401/command
REMOTE_MODULE command /usr/bin/uptime
PUT /tmp/tmpCAU0Pt TO /root/.ansible/tmp/ansible-1399929485.49-54151257537199/command
EXEC /bin/sh -c ‘/usr/bin/python /root/.ansible/tmp/ansible-1399929485.52-91011211544401/command; rm -rf /root/.ansible/tmp/ansible-1399929485.52-91011211544401/ >/dev/null 2>&1’
EXEC /bin/sh -c ‘/usr/bin/python /root/.ansible/tmp/ansible-1399929485.49-54151257537199/command; rm -rf /root/.ansible/tmp/ansible-1399929485.49-54151257537199/ >/dev/null 2>&1’
EXEC /bin/sh -c ‘/usr/bin/python /root/.ansible/tmp/ansible-1399929485.48-193948231594553/command; rm -rf /root/.ansible/tmp/ansible-1399929485.48-193948231594553/ >/dev/null 2>&1’
EXEC /bin/sh -c ‘mkdir -p $HOME/.ansible/tmp/ansible-1399929485.52-98282571951561 && echo $HOME/.ansible/tmp/ansible-1399929485.52-98282571951561’
REMOTE_MODULE command /usr/bin/uptime
PUT /tmp/tmpqfWmwp TO /root/.ansible/tmp/ansible-1399929485.52-98282571951561/command
EXEC /bin/sh -c ‘/usr/bin/python /root/.ansible/tmp/ansible-1399929485.52-98282571951561/command; rm -rf /root/.ansible/tmp/ansible-1399929485.52-98282571951561/ >/dev/null 2>&1’

I’m running ansible 1.4.5 (Ubuntu 10.04 starting point, seen above as host4)
Even running the same command on one of the hosts (doesnt matter which one) just stalls … (even after checking on the destination and verifying that the command was finished… All that remains is the “sshd: root@notty” trace…

Am i doing it wrong?

Thanks!

2

To complete the debug info:

After a CTRL-C on the console i get the following output:

Traceback (most recent call last):
File “/usr/lib/pymodules/python2.6/ansible/runner/connection_plugins/paramiko_ssh.py”, line 321, in close
self.ssh._host_keys.update(self.ssh._system_host_keys)
File “/usr/lib/python2.6/UserDict.py”, line 155, in update
for k, v in other.iteritems():
File “/usr/lib/python2.6/UserDict.py”, line 110, in iteritems
yield (k, self[k])
File “/usr/lib/pymodules/python2.6/paramiko/hostkeys.py”, line 258, in getitem
ret = self.lookup(key)
File “/usr/lib/pymodules/python2.6/paramiko/hostkeys.py”, line 224, in lookup
if (h.startswith(‘|1|’) and (self.hash_host(hostname, h) == h)) or (h == hostname):
File “/usr/lib/pymodules/python2.6/paramiko/hostkeys.py”, line 312, in hash_host
hmac = HMAC.HMAC(salt, hostname, SHA).digest()
File “/usr/lib/python2.6/dist-packages/Crypto/Hash/HMAC.py”, line 56, in init
self.inner.update(_strxor(key, ipad))
File “/usr/lib/python2.6/dist-packages/Crypto/Hash/HMAC.py”, line 16, in _strxor
return “”.join(map(lambda x, y: chr(ord(x) ^ ord(y)), s1, s2))
File “/usr/lib/python2.6/dist-packages/Crypto/Hash/HMAC.py”, line 16, in
return “”.join(map(lambda x, y: chr(ord(x) ^ ord(y)), s1, s2))
KeyboardInterrupt
host3 | success | rc=0 >>
host3

(where the command starting this was ‘ansible host3 -vvvv -m shell -a “hostname”’

Is the python 2.6 the problem here?

3

If you have the playbook configured to sudo and you didn’t pass --ask-sudo-pass, that may be the problem.

We have a ticket open to irmprove the detection of a sudo prompt.

4

FYI, this question in the future is more apt for ansible-project, as it’s about usage, rather than developing ansible itself.

Not a problem but wanted to point that out.

Thanks!

5

Michael,

Thanks for the reponse. I realized after posting the opening post that it was the developers list… Sorry, my bad, my next post will be placed in the correct list :slight_smile:

I’ve also haven’t been able to reproduce on Ubuntu 14.04 (ansible 1.5.4) so it seems it might just be an old version not working a bit wonky…
Thanks again for the prompt reply. Seems upgrading to a more recent (non-apt due to 10.04 host) might solve the issue?

6

I’m not aware of any recent hanging scenarios in any version, FWIW – except in cases where a command might have gone interactive (depends on the command).

7

Well as said, the command example requires no sudo, isnt particularly
heavy, just returns the hostname :slight_smile: (also tried "-m ping" to no avail)

Thus all the following examples exhibit the same behaviour:
ansible office -vvvvv -a "uptime"
ansible office -vvvvv -a "hostname"
ansible office -vvvvv -a "w"
ansible office -vvvvv -m ping

control host: Ubuntu 10.04
ansible --version: ansible 1.4.5
python --version: Python 2.6.5
Met vriendelijke groet,

Robin Speekenbrink

8

Please paste the full command line you are running (including the shell prompt) and the full output you get (up to and including the next shell prompt).

Thanks!

9

Here you go, doesnt differ much from my opening post:
root@host4:~# ansible office -vvvv -a "uptime"
<host1> ESTABLISH CONNECTION FOR USER: root on PORT 22 TO host1
<host2> ESTABLISH CONNECTION FOR USER: root on PORT 22 TO host2
<host4> ESTABLISH CONNECTION FOR USER: root on PORT 22 TO host4
<host3> ESTABLISH CONNECTION FOR USER: root on PORT 22 TO host3
<host1> EXEC /bin/sh -c 'mkdir -p
$HOME/.ansible/tmp/ansible-1400077153.93-230744874133396 && echo
$HOME/.ansible/tmp/ansible-1400077153.93-230744874133396'
<host1> REMOTE_MODULE command uptime
<host1> PUT /tmp/tmpHXoNFe TO
/root/.ansible/tmp/ansible-1400077153.93-230744874133396/command
<host3> EXEC /bin/sh -c 'mkdir -p
$HOME/.ansible/tmp/ansible-1400077153.91-47758668405338 && echo
$HOME/.ansible/tmp/ansible-1400077153.91-47758668405338'
<host2> EXEC /bin/sh -c 'mkdir -p
$HOME/.ansible/tmp/ansible-1400077154.0-155050846726559 && echo
$HOME/.ansible/tmp/ansible-1400077154.0-155050846726559'
<host1> EXEC /bin/sh -c '/usr/bin/python
/root/.ansible/tmp/ansible-1400077153.93-230744874133396/command; rm
-rf /root/.ansible/tmp/ansible-1400077153.93-230744874133396/

/dev/null 2>&1'

<host3> REMOTE_MODULE command uptime
<host3> PUT /tmp/tmpHXoNFe TO
/root/.ansible/tmp/ansible-1400077153.91-47758668405338/command
<host2> REMOTE_MODULE command uptime
<host2> PUT /tmp/tmpRSTQ3_ TO
/root/.ansible/tmp/ansible-1400077154.0-155050846726559/command
<host2> EXEC /bin/sh -c '/usr/bin/python
/root/.ansible/tmp/ansible-1400077154.0-155050846726559/command; rm
-rf /root/.ansible/tmp/ansible-1400077154.0-155050846726559/

/dev/null 2>&1'

<host3> EXEC /bin/sh -c '/usr/bin/python
/root/.ansible/tmp/ansible-1400077153.91-47758668405338/command; rm
-rf /root/.ansible/tmp/ansible-1400077153.91-47758668405338/

/dev/null 2>&1'

<host4> EXEC /bin/sh -c 'mkdir -p
$HOME/.ansible/tmp/ansible-1400077154.08-121377164394763 && echo
$HOME/.ansible/tmp/ansible-1400077154.08-121377164394763'
<host4> REMOTE_MODULE command uptime
<host4> PUT /tmp/tmpHXoNFe TO
/root/.ansible/tmp/ansible-1400077154.08-121377164394763/command
<host4> EXEC /bin/sh -c '/usr/bin/python
/root/.ansible/tmp/ansible-1400077154.08-121377164394763/command; rm
-rf /root/.ansible/tmp/ansible-1400077154.08-121377164394763/

/dev/null 2>&1'

^C32
ERROR: interrupted

After a few minutes i CTRL-C (thus ^C32 and the error) otherwise the
prompt would never be returned

Running the uptime command on a single host (out of the office group)
through strace:
strace ansible host3 -vvvv -a "uptime"
[... alot ... ]
but ends in
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAIT_PRIVATE, 0, NULL) = -1 EAGAIN (Resource
temporarily unavailable)
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAIT_PRIVATE, 0, NULL) = -1 EAGAIN (Resource
temporarily unavailable)
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0xdd3540, FUTEX_WAKE_PRIVATE, 1) = 1
repeating over and over again...

Hope this helps!
Met vriendelijke groet,

Robin Speekenbrink