Andy16 wrote:silvertip257 wrote:
Without your script, it's a bit hard to know for sure (but here goes my guess below).
I think I missed the point - sorry for that.... and thanks for the fast reply!
script is doing quite a lot in logon, preprocessing and so on and cannot be run in parallel (only single user access)
Result is an output like
data1:123 data2:234 data3:345 data4:435 data5:354 data6:782 .............
In graph A I wanna show data1 and data2,
in graph B only data3 and data4
in graph C only data1 and data5 and data6
So script shall run once every 5 min, update one (or several) rrd files and graphs shall read data from the file(s)
How can I set that up?
thanks for your help
I would configure your script so you pass a parameter (input string
[0]) to tell your script which pieces of data to retrieve and print to stdout (for your Data Input Methods).
You would end up running your script 3 times, but it would retrieve smaller subsets of data.
Ex:
Input: a ... Output: data1:123 data2:234
Input: b ... Output: data3:345 data4:435
Input: c ... Output: data1:123 data5:354 data6:782
Or potentially have 4 Data Input Methods (and Data Templates) data1, data2, data3+data4, data5+data6
Data Input Method > Data Template > Data Source Item
(a single DT can contain multiple DSIs and a single RRA file is created per DT)
[ I checked and haven't done what you describe above... ]
Either way I suspect you will end up with duplicate data in different RRAs (with the 3 part method) regarding data1.
While even if you had data1 in its own DT (4 part method), I don't believe Cacti is smart enough to skip creating another data1 Data Source. This behavior would likely happen if you're creating Graph Templates, but not for one-off graphs where you reference existing Data Sources.
[0]
http://docs.cacti.net/manual:087:3a_adv ... ut_methods
grrr - so no way having a huge poller script and feed multipe graphs??? Graph : poller is a 1:1 relation?