Ad blocker detected: Our website is made possible by displaying online advertisements to our visitors. Please consider supporting us by disabling your ad blocker on our website.
See how it automatically defines a heap more data sources for AVERAGE,MIN,MAX,LAST rather than just AVERAGE on yours. Thats what the problem is. To fix it on yours you need to adjust the CDEF to suit only have 3 data sources instead of the 12.
So instead of the CDEF being a,e,+ you will probably want a,b,+
Does that make sense?
So does this mean .8.7.d treats the datasources differently? <groan>...
I have the same 3 datasoureces, instead of the 12 you posted...
No longer a n00by, probably, by now an 0ldby
Now Head of Technology at RSCH, back to the prickly subject of Monitorring....
So I figured out half my problem. I was trying this out on a PPC build (on a mac mini) and, for some reason, the wmi-client doesn't work. I installed wmi-client on a x86 system and am able to run successful queries... on Windows 2003/XP systems. Still no dice with Windows 2008.
I'm using 0.8.7b but we had the same problem going from 0.8.7 -> 0.8.7b where it changed from just using AVERAGE to using the 4 inputs per data source. In face it broke a huge number of graphs for us when we upgraded. If its changed again in 0.8.7d it's going to be a right pain to upgrade again
argon0 wrote:Same problem with Memory Usage graph.. This time with e and i.
Does this mean we need to check all our cdefs? To make sure they don't point to wrong data points?
Claymen If you post up the debug from the WMI graphs that use the Cdefs you defined I'll compare them to my install..
BTW are you using .8.7d or .8.7b?
Argon0
Basically any CDEF I have made will be broken because of this. I'm not running 0.8.7d so I can't see exactly what is happening. From my understanding the reason they added the extra inputs in on the graph was to improve the accuracy but as you can see at the expense of breaking CDEF's. I'll see if there is a way to work around it but I somewhat doubt it, I couldn't find anywhere in cacti to even configure this
I got bored over the weekend (whilst recovering from a concert) and setup MantisBT to track bugs. If you could submit bugs to that so I can track them better that'd be awesome.
I shall surely do so if I come across other bugs, but it seems pretty good apart from that...
Would be good to automate the data discovery (e.g. Storage Groups, etc...) on the Exchange Server, but I can live without it. I shall also (be trying to) write a graph template to graph all SGs on one graph for my site.
BTW do you know if there is a _Total namespace for each Storage Group, rather than for all DBs, and each individual DB?
As an aside, I noticed that my graphs kept breaking, and the wmi query wouldn't work remotely (error : NTSTATUS: NT code 0xc0041010 - NT code 0xc0041010), but it did locally (thanks to Ben Coleman's WMI Query tool), so I ran the wmiadap /f again, and it started working again... <huh>
No longer a n00by, probably, by now an 0ldby
Now Head of Technology at RSCH, back to the prickly subject of Monitorring....
To do data discovery it needs to be a data query and for that it may be better long term to look at the plugin arch. But at the same time it requires the plugin arch which is a pain to maintain.
I will look at writing a data query for things like disks and storage groups etc, things where it makes sense. It is something I have wanted to do but just needed something in place first.
Not sure about a _Total for each storage group, only for the databases themselves. If you come across one let us know as it'd probably useful for others
That's a little odd about wmiadap /f fixing the remote query when it worked locally... But then again it does refresh all the counters. The biggest problem we have is the firewall blocking RPC traffic. We do allow it but we set the dynamic RPC ports to a specific range and only allow that range.
And here is a variant of the Database Performance Graph Template - widened it to 750 pixels, and changed Gprint type for legends from Exact Numbers to Normal..
Prior to this I was getting 0 for all entries, now I get a an entry (OK each one is x milli, but still an entry, which makes more sense imho )
Don't know if I'm just missing this or whether you just CANNOT do this, except via Graph Templates...
Would like to be able to change the Exchange Active Connections graph to name the databases, instead of having it just say Database1, Database2.... on the graph....
No longer a n00by, probably, by now an 0ldby
Now Head of Technology at RSCH, back to the prickly subject of Monitorring....
Don't know if I'm just missing this or whether you just CANNOT do this, except via Graph Templates...
Would like to be able to change the Exchange Active Connections graph to name the databases, instead of having it just say Database1, Database2.... on the graph....
And is there an easy way to copy several lines of a graph then edit them, rather then adding each line one at a time (I remember something - a bit of a hack I did - from ages ago where you could add 3 or 4 lines at once...)
No longer a n00by, probably, by now an 0ldby
Now Head of Technology at RSCH, back to the prickly subject of Monitorring....
And it can integrate with SVN!