I have a question regarding the snmpWalk console example. I have implemented it into our SNMP manager, but I have benn having issues getting all the expected values returned. Sometimes I’ll get 2100+ results while other times I’ll get around 1400. I have more constant results when having a lower bulk max value though. My question is how can i return more consistent results with a higher bulk max value? Would this possibly be tied to the timeout value timing out before the bulk max has returned all items?
I have tried to increase the timeout duration without much success. I was wondering if you all had any thoughts or insignts on the matter.
What kind of error do you get on intermediate requests when you receive less than the expected amount of variables?
If you do not get any error, the agent (command responder) most likely has a bug regarding:
lexicographic ordering
Returning GETBULK results before memory resources are exhausted
That’s a good question, I let you know next time I have a chance to investigate. Last time I had a chance I decided to tweak the get_bulk method’s arguments with a Bulk_max value of 100 and the non-repeaters to be also 100. After doing so it appeared as though I got the full 2100+ results. However, I was surprised that the performance of the walk didn’t improved, if anything it got worse …
To be honest I really don’t understand the use of non-repeaters either. Supposedly it’s to represent the number of unique elements per request? Since I’m attempting to walk a 9*250 table I’m assuming they will all be non-repeaters?
Non-repeaters are for scalar values. The value specifies how many OIDs from the beginning of the GETBULK variable binding list should be handled as GETNEXT by the agent.
Thus, if you want to get a table, the non-repeaters should be 0.
Just a small addition to max-repetitions: The value is the maximum amount of repetitions. The agent is allowed to reduce this value. So there is nothing wrong with this:
Sometimes I’ll get 2100+ results while other times I’ll get around 1400.
