Ansible exit statuses

I need the exit codes for Ansible for reporting purposes.

The exit status or return code of a process in computer programming is a small number passed from a child process (or callee) to a parent process (or caller) when it has finished executing a specific procedure or delegated task.
wikipedia.org/wiki/Exit_status

Google can’t seem to show me information about the Ansible commandline exit statuses though. So I looked them up & here they are. The exit states were different in older versions, only the more recent versions are compliant with standard Linux exit behaviour.

Ansible exit 0
OK
This includes “no hosts matched”.

Ansible exit 1
Something’s wrong.
Includes wrong options, missing arguments, errors in the playbook, ansible specific errors.

Ansible exit 2
Failed hosts

Ansible exit 3
Dark hosts

source: source code VERSION 1.7, bin/ansible file.

This should be in the Docs & if I ever figure out where to submit this & what hoops to jump through to get this in there, I will. But first some more work! And in the mean time, at least it’ll be on here..

Ansible-pull unknown hostkey.

I encountered an Ansible-pull “unknown hostkey” error.

Starting ansible-pull at 2014-06-03 11:17:00
localhost | FAILED >> {
    "failed": true, 
    "msg": "git.repo.com has an unknown hostkey. Set accept_hostkey to True or manually add the hostkey prior to running the git module"
}

After some digging around, turns out the latest Ansible has no such problem, but the installed “ansible 1.5.5″ running the pull has. Apparently that used to be a bug.

Simple solution is obviously to upgrade, but why go for the simple solution. Turns out this ignores most of the options concerning host-keys, except /etc/ssh/ssh_known_hosts

That means that adding the repository ssh-key will enable you to pull the repository.

ssh-keyscan git.repo.com >> /etc/ssh/ssh_known_hosts

Fun fun.

Atlassian stack Frustrations

Most of these are rethorical questions since I know that most of the reasons are design flaws, oversight, money grab and similar constraints. Setting it for auto post in about a week!

Why doesn’t Jira & Confluence integrate better? Like decent issue widgets without having to manually create rss exports in Jira. They feel like 2 separate products with some very very basic integration!

Why can’t I have all Confluence emails as notifications in their interface? How can I enable more notifications types?!

why doesn’t the Jira Kanban board automatically refresh?

Why does it take so bloody long to load the Jira Kanban board and whats with that spinner once it’s half loaded?!

Why aren’t the Jira/Confluence pages responsive?? (ie responsive web design)

Who ever thought it a good idea to display comments chronologically on the issue page (I like) and reverse chronologocally on the Kanban board issue zoom?! (Yes I get that that way & with a bit of luck we’ll always see the most recent comment first, but just no.)

Why do some pages not decently load, but stay blank instead? I mean, what has to be going on on that page for shit like that to happen?!

H.R. Giger died today.

A sad day.

Joining in on the chorus of sad Giger fans.

It’s sad news, I wouldn’t have minded meeting the man some day and I’ve been planning on buying Giger art for a while now, but I’m guessing they will become unaffordable instead of just expensive, soon.

I’m still very happy I went to see his museum in Gruyeres a few years back. Very powerful. An impressive artist. Who created art that radiates emotion off the canvasses. I still remember Viona being impressed by the detail and craft in the enormous wall covering canvasses and her reaction and emotion in the smaller red room.

Sad news.

 

This seems like a very interesting interview. Very personal with museum footage etc.

Reading “Daemon”

Daemon. By Daniel Suarez.
Interesting so far but it seems tech’d-up.

All the techno babble is ok but nobody looks at hardware and says “it has a TCP/IP stack that enables to to interact with IP networks over this cat5 cable”. Either the people you’re talking to know what you’re talking about and you’d just say talk about a network interface & cable. Or they don’t and then they would understand even less of what your saying than without those words.

So it’s just technical correct verbosity to up the tech appeal and credibility for non tech people.

(Originally a Goodreads update but apparently those are limited to 420 characters. And that’s a weird short limit for a site about books.)

On a train, not entirely sure the doors will open when I get to where I’m going.

On a train. Again.

Last 3 wagons will not open where I’m going. When asking the conductor dude seconds after the automated message, he first looked befuddled. Then, after asking again, if this wagon was ok to get off in my destination, he looked again slightly confused, then barely restrained himself from shaking the clouds out of his head and said, yes, yes, this wagon will do for your destination.

This leaves me wondering. Did he take into account what the message just announced? Or did he just figure “yes, this train stops at your destination, can’t you read, it’s right up on the screen”. In all honesty, I did shake him by not needing to get by him to the bathroom and ask him questions instead.

Seems to be having a good time with those ladies though.

I’m supposed to be working now, damned!
Don’t waste my time, train!!

Although time was already wasted, the train was 5 minutes late and although considering those other trains it was late for and the fact that they were already up on the board, they didn’t mark it as 5 minutes late about 4 minutes after the planned time of departure.
I get the effort to improve delays on the board, but not marking them is just abusive.
If I had known I’d have had that coffee sitting down & working instead of dodging people trying to get to those 2 late trains that delayed mine.

aah yes.
A first in many rants to come, I fear.

If the absolutely flaky internet will allow me. The plan of working on the train will be more difficult than ever conceived.

Transcendental shower time

With SPEKTRMODULE.

..
YOGA. But without the moves.

Holistic.
5 track nitenitemusic.
Incense burner.

Visualising the perfect house. A bathroom with an outside door. And a location that supports howling at the stars at night. Naked & wet. Without neighbours noticing. Sauna. Plants blanketing the horizon for serenity and scents. In the woods?

When the fog lifts

image

Ps. SPEKTRMODULE 17 Revisitation.

http://spkmdl.libsyn.com/spektrmodule-17-revisitation

Het Depot, Leuven.

Gisteren avond was het Me First & the Gimme Gimmes in Het Depot in Leuven. Voorafgegaan door Old Man Markley. Beiden zeer goeie bands, entertainers. Met als grote verrassing toch wel Old Man Markley, waar we dan direct de obligate CDs van aan de collectie toe voegden.

Doch mijn voornaamste verbazing dan al voorbij was. Ik was namelijk nog nooit in Het Depot geweest en hoe het mij voorgesteld was door een niet nader genoemde vriend, verwachtte ik me aan een oude hangar met een herbestemd laadplatform als afgeleefd betonnen podium. Er was ons oorschade van de ergste orde beloofd en ik had Viona al gewaarschuwd dat om zich door de geluidsbarrage te worstelen, ze zeker oordoppen in zou moeten hebben!

Stel je dan even mijn verbazing voor bij het betreden van het hypermoderne gebouw dat eerder voor toneel dan concerten ingericht is. Cinema stijl zitplaatsen en een par terre voor wie het wat wilder wilt. En het was uitverkocht, dus het ging er feestelijk aan toe.

image

En wat met het geluid? Gemiddeld 101db over de hele lijn. Dat weet ik, want ik stond in de buurt van de PA waar dat mooi groot te lezen stond op een groot scherm. Niveau tussen de 90 en 110, dus ook daar alles behalve te klagen.

image

Zeker een plek waar in nog naartoe wil! (Behalve dan een beetje ver)
En een heel mooie avond. Aanrader!

ps. What’s with bands not having proper websites anymore these days?!

Debian Squeeze to Wheezy dist-upgrade breaks sasl+mysql

Upgrading broke my postfix+sasl+mysql auth, cripling my SMTP service. Because libsasl2-2 that was upgraded from 2.1.23 to 2.1.25.

I upgraded Debian Squeeze to Wheezy and most of it worked within acceptable parameters. There were problems, but nothing big, except that it killed my SMTP service. Upgrades, the gift that keeps on giving.
Many people offered that downgrading would fix things, but that can’t be the best solution. So I didn’t.

The update triggered the following error:

SASL LOGIN authentication failed: no mechanism available

This was fixed by commenting the “auxprop_plugin: mysql” line, ie not specifying the plugin. Thanks Marcel.

That yielded the following errors

sql plugin couldn't connect to any host
warning: SASL PLAIN authentication failed: generic failure

(gotta love “generic failure”. Generic: “characteristic of or relating to a class or group of things; not specific.” Not being able to connect to the database isn’t what I’d report as a “generic failure”)

This is where it gets a bit ludicrous. There is no changelog around and google didn’t help much either, until I got to reading the very long thread on the Ubuntu launchpad where “dbileck” mentioned sql_passw changing to sql_passwd. So a parameter changing, adding a “d” and breaking all reverse compatability.
Now, in all honesty, http://www.postfix.org/SASL_README.html#auxprop_sql has the correct parameter & not the old sql_passw, but a small note, mentioning the extra “d” would’ve helped a lot!

In the end, these are the changes to /etc/postfix/sasl/smtpd.conf that fixed it.

- auxprop_plugin: mysql
- sql_passw: XXX
+ sql_passwd: XXX

Oh, the pleasure of introducing small undocumented changes in config files.

And now I’m a happy camper again, my network scanner is able to mail documents again!

Edit: haha, a bit of a brain bubble, instead of Squeeze, I wrote Jesse. I’m guessing a short moment of confusion because I run “Jessie”, the current testing version on my laptop! No worries or weirdness though, a “simple” case of upgrade from Squeeze to Wheezy.