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.
I would like to use the first integer as the value for my ds. I tried many different approach su as using VALUE/REGEXP in my <source>Â tag and I can't get it to work.
Poller[0] Host[5] DS[122] WARNING: Result from SNMP not valid. Partial Result: 55C (131
Poller[0] Host[5] DS[123] WARNING: Result from SNMP not valid. Partial Result: 50C (122
Poller[0] Host[5] DS[123] WARNING: Result from SNMP not valid. Partial Result: 23C (73
Poller[0] Host[5] DS[123] WARNING: Result from SNMP not valid. Partial Result: 55C (131
Poller[0] Host[5] DS[124] WARNING: Result from SNMP not valid. Partial Result: 50C (122
Poller[0] Host[5] DS[124] WARNING: Result from SNMP not valid. Partial Result: 26C (78
Poller[0] Host[5] DS[124] WARNING: Result from SNMP not valid. Partial Result: 55C (131
Poller[0] Host[5] DS[125] WARNING: Result from SNMP not valid. Partial Result: 50C (122
Poller[0] Host[5] DS[125] WARNING: Result from SNMP not valid. Partial Result: 26C (78
From what I've read, it is possible that VALUE/REGEXP cannot be used for an output field. However, the cacti documentation says it can be used.
The VALUE/REGEXP feature doesn't work with the direction set to output, I've read through the sourcefiles and found the proof that it only works with the direction set to input.
From lib/data_query.php: (the only file that contains the string 'VALUE/REGEXP')
My first fix for this was a php script to pull the point and return (print dsname:value) it as a data point, using the string manipulations in php. I feel like this is very SLOW. I'd like to redo this in the cacti scripting engine ( where PHP and DB hooks exist already ) and see what sort of improvement can be had there. ( Reading on that tonight. )
I am planning to rewrite the php script so it handles all the string OID's from a host type and returns them en-mass to one RRD file, to reduce php startup overhead; walk the OIDs in a batch/get table and then manipulate the results and return them all at once. This would be more efficient, process wise but also lock all the data points to the same refresh/get rate.
My final option is to recompile spine with the no_snmp_string_return and see what spine returns from those points. This might give you both numerics mushed together, guessing from your snmpwalk results pasted above.