On mån, 2008-05-12 at 11:18 +0530, Rupa P V wrote: > Thanks a lot for the reply. But I have a doubt : how to generate > template code for this "basic table helper", if not using mib2c ?
Sadly the answer to this question is "With a text editor". > What is the command to generate this code framework for my mibnode ? There is none. Contributions are welcome. I would like to reiterate that the first thing you have to decide is how getting to the next key value is to be done, once that is finished the mechanics of instrumenting the code is rather simple and you should be able to use one of the existing helpers. /MF > -rupa. > > On 5/12/08, Magnus Fromreide <[EMAIL PROTECTED]> wrote: > > On mån, 2008-05-12 at 09:36 +0530, Rupa P V wrote: > >> I am trying to find the correct mib2c conf file for implemeting the > >> low level snmp routines for my database. These are my requirements: > >> > >> 1) Given the index/indices in the OID extracted from the SNMP Varbind, > >> I can directly access the node which has the required info. I don't > >> want to go through a getfirst-getnext cycle to find the correct row > >> for an SNMP GET request. My database is maintained as a hash table and > >> so I can access my data by hashing on my indices. > >> > >> 2)My data base is not sorted, so for SNMP GetNexts , I dont mind > >> getfirst-getnext cycles being called. > > > > This is problematic. > > > > For GET and SET you should be able to use the basic table helper (not an > > mib2c template, I mean the helper!) > > > > The problem is with GETNEXT, you are given an index value and are then > > expected to return either the first entry with key > index or the first > > entry with key >= index, now given that your data are stored in a hash > > table you need to somehow store the information of where to find the > > next item. Thus the optins I can see are > > > > * Keep an ordered index over the existing keys, either in a > > separate table or as part of the objects of the first table. > > (cost: memory, complexity) > > A variation of this is to change the dataase to an ordered data > > structure (cost: depends on the data structure, complexity) > > > > * Make exhaustive search of the index set. (cost: performance) > > > > /MF > > > >> I tried mfd (unsorted external also) and iterator options , but > >> iterator iterates through the getfirst-getnext cycles for a single > >> SNMP GET itself. If mfd is used, then there is memory required for > >> cache. > >> > >> Is there a method which meets these requirements without iteration and > >> cache ? I am using netsnmp 5.4.1. > >> Plz let me know if anyone knows the answer. > > > > > > ------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone _______________________________________________ Net-snmp-coders mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/net-snmp-coders
