Ad Widget

Collapse

Best practise with long tests.

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • akapap
    Member
    • Nov 2009
    • 35

    #1

    Best practise with long tests.

    Good morning,

    I've been thinking about difference with passive and active tests. I still have some doubts which should go with long, rather predictive tests?

    On the machine where I have clamav-milter running I have installed zabbix's agent. With perl code I am testing milter antyvirus software. In the 98% I get the answer from the perl code within 2 seconds. So in passive tests I look into the file where I write answer and use my own key. But:

    - when the file have no data, zabbix states that tests is unsupported,
    - when the perl agent crash for all the future time in the file there will old data.

    So maybe the active tests with the trigger nodata is my solution? When I won't get any data from the hosts I will alarm. But there is a question if active tests are prepared for regular (every 1 minutes) testing? For every time I will be have to raise a binary with zabbix_sender and still I will have to prevent double running the same script (is one stays longer then 60s).

    So, what should I use. The passive tests, I think are for fast, reliable, testing. So everything else should go for active with nodata trigger?

    Best regards,
    E.
Working...