email notifications are not being delivered

[mmestnik]

One thing I've done in testing is to convert line endings in most of the files.
Note: Take special care to enter the '^M' properly, in most shells this is done with the following key sequence. Ctrl-v Ctrl-m or simply Ctrl-vm.

Code: Select all

find /usr/local -type f -exec grep -Z --binary-files=without-match -l '^M' '{}' + | xargs -0 -n1 dos2unix -k

...However this was found not to help, BTW.

[mmestnik]

Here is one tracker item that may be of interest.
http://tracker.nagios.com/view.php?id=28

[mmestnik]

Currently we believe that this issue can be corrected by applying the following patch. This can be run as a script from the command line. Unfortunately PHP-BB has a sleek way of mangling the provided content, so I've also supplied a compressed base64 encoded copy that you can copy/paste.

Code: Select all

/usr/bin/patch /usr/local/nagiosxi/cron/eventman.php <<EOF
RCS file: /usr/local/nagiosxi/cron/eventman.php,v
retrieving revision 1.1
retrieving revision 1.2
diff -u -r1.1 -r1.2
--- /usr/local/nagiosxi/cron/eventman.php       2010/04/01 18:02:12     1.1
+++ /usr/local/nagiosxi/cron/eventman.php       2010/04/01 18:02:38     1.2
@@ -9,7 +9,9 @@
 //define("BACKEND",1);

 require_once(dirname(__FILE__).'/../html/config.inc.php');
-require_once(dirname(__FILE__).'/../html/includes/utils.inc.php');
+//require_once(dirname(__FILE__).'/../html/includes/utils.inc.php');
+//require_once(dirname(__FILE__).'/../html/includes/components.inc.php');
+require_once(dirname(__FILE__).'/../html/includes/common.inc.php');

 $max_time=59;

EOF

Code: Select all

tr -d '\n' <<EOF | base64 -d | bzip2 -d | patch /usr/local/nagiosxi/cron/eventman.php
QlpoOTFBWSZTWb1toAsAACvfgAAyVO/++n8tAACv5/9AMAEmtrQklKfpT9MoeqeKbRqaMQxPUaZ6
oaNAEkqY0TQNAAaA0NAA0GglKm9EnqeUyADIAAANDQK7uRSHa3WyKdCgMmaAXvotsgdjtdsciLkb
0jBF0WDERVTZcZkOhzZkXoxoifHNToOCJIjI1moqKKCR5IgwKSoMHlUlFMZrIh0DToIhtYdG0aUs
bn7XlGe3dtJF5oYNPgWd2GrCfGaNiOKOE0TcIZHBFpFMxOMihMj2MgmF+CJD0CHIadFFI1UFBwsO
VJ3RGoSnv8nRBgL4LSFdCwg8jIEgRn2BD7yLsLD4xCW0OQ8hkMSLwRkTjAi46o/UrUnwRyKUbN8j
1ciJIsgGUsGK7ytDPClFcIi9IfxdyRThQkL1toAs
EOF

[gary_ford]

reference the above - could you be more specific on what i need to do, you are assuming I have a brain. (do i simply "select all code" from the second section and paste into putty session?)

does 1.1h fix this issue?

[gary_ford]

ok i did that and it fixed it immediately. hopefully this is included in the next release?

I dont dare to think what would have happened should that have been broken in a production release. I hope more stringent testing will be put in place for future releases..

[mmestnik]

Yes, 1.1G was a must have this bug fixed now release that caused a second this other bug must be fixed now release, 1.1H.

We are working on building a huge regression testing environment and script. It might not be functional for the next two releases or so, but we consider it vary important.

[gary_ford]

just to be clear 1.1H did NOT fix the email issue, the code you suppied fixed it. I tested after 1.1H and it was still an issue - i then used the code and that fixed the error so as far as I am aware 1.1H is broken also.. (or maybe it was 1.1g that broke it and if you go straight to that it will be ok.

Testing is the biggest issue so far for me - you need to have more stable releases to make this workable. The company I am working for will 90% buy the product soon but we are in testing at the moment with nagiosxi and about to deploy to 25% of our estate, should the product fail again with the 25% being monitored then i can see it being pulled.

Please do more rigorous testing in the future before releasing code. I would rather wait an extra month for a minor bug fix than it be rushed and break major functionality.

[mmestnik]

One quote that's been mentioned around the office, though it's vary much overshadowed by the"We need to make sure releases are working more then we need releases to go out," is that the industry has already been programmed to not upgrade. I might go so far as to suggest for the time being you should treat our upgrades like you would an OS service pack from Microsoft. If I'd be a gambling man it would be a sure bet that you have a box dedicated to testing these, no?

With NagiosXI it would be trivial for you to maintain a VM some where to see if our upgrades will work in your environment.

Some times fixing a bug is more critical then the perceived potential to create more, though once we have an automated regression test in place it will no doubt be mandatory for every release regardless of critically.

[gary_ford]

I will be more diligant in my testing going forward, and snap before i do so - this is for sure. I had taken my eye of the ball as this started as a proof of concept for me but has grown into a solution that is certainly workable and considering the hours already invested I would be silly not to pay it with a little more respect in the testing dept.

Just as long as you do also

[mmestnik]

That's something I can say we will do for sure.