About net-snmp table iterator performance

Hi all,

In the net-snmp website. I find following description : "Through the
get_first_data_point and get_next_data_point hooks,
the table_iterator helper will repeatedly call your hooks to find
the "proper" row of data that needs processing."  Then I
had write a simple example to test the way table_iterator helper working, in
my example I had create a link-list of 200 nodes
to simulate the table. Every node is a struct like below:

*/*My table name is ggsnPicStatsTable, and has three colums which are
ggsnPicIndex, ggsnPicAddress and ggsnPicNbrOfActivePdpContexts*/*
*            *
*             typedef struct ggsnPicStatsTable_full_s       *
*            {*
*                 int    ggsnPicIndex;     /*In my code, it's value is from 1
to 200*/*
*                 int    ggsnPicAddress;     /*In my code, it's value
is 2,4,6,...400,
double the ggsnPicIndex*/ *
*                 int    ggsnPicNbrOfActivePdpContexts;     /*In my code,
it's value is 3,6,9,...900, treble the ggsnPicIndex*/*
*                 struct ggsnPicStatsTable_full_s *next;*
*            }ggsnPicStatsTable_full_t;*
*
*
*Now I start the master agent and my sub-agent, and use some snmp command to
test the way table_iterator helper working, *
*in other words, to test how the function example_table_get_first_data_point
and example_table_get_next_data_point are called. *
*To test this, I add some printf lines in the function
example_table_get_next_data_point,
as below:*
*
*
*
            static netsnmp_variable_list *
            example_table_get_next_data_point (

void **my_loop_context,

void **my_data_context,

netsnmp_variable_list *put_index_data,

netsnmp_iterator_info *mydata__unused)
            {

.......

printf("%d\n",entry->ggsnPicIndex)

........
            }
*
*
*
*And there are some detailed results of my testing shown as below:*
*    1. When I input command "snmptable -v 2c -c public localhost
GGSN-MIB::ggsnPicStatsTable", the output is : *
*       *
*                           ggsnPicAddress
 ggsnPicNbrOfActivePdpContexts*
*           Wrong Type (should be IpAddress): 2       Wrong Type (should be
Gauge32 or Unsigned32): 3*
*           Wrong Type (should be IpAddress): 4       Wrong Type (should be
Gauge32 or Unsigned32): 6*
*           Wrong Type (should be IpAddress): 6       Wrong Type (should be
Gauge32 or Unsigned32): 9*
*           ..........*
*           Wrong Type (should be IpAddress): 400       Wrong Type (should
be Gauge32 or Unsigned32): 600*
*
*
*       you can ignore the warning"Wrong Type (should be IpAddress)", and
just focus the output value which is correct.*
*
*
*       On the other hand the output of my printf in the function
example_table_get_next_data_point  is :
*
*

*
*

 1,2,3,4,5,6,....199,200,1,2,3,....199,200,1,2,3,....,199,200.......1,2,3....199,200,
from
1 to 200, and repeated 200 times.
*
*

*
*

*
*
     2. When I input command "snmpbulkget -v 2c -c public -C n0 -C r50localhost
GGSN-M
IB::ggsnPicIndex
GGSN-M
IB::ggsnPicAddress GGSN-MIB::ggsnPicNbrOfActivePdpContexts",
*
*
         the output is :
*
*

*
*
            GGSN-MIB::ggsnPicIndex.1 = INTEGER: 1
*
*            GGSN-MIB::ggsnPicAddress.1 = Wrong Type (should be IpAddress):
INTEGER: 2*
*            GGSN-MIB::ggsnPicNbrOfActivePdpContexts.1 = Wrong Type (should
be Gauge32 or Unsigned32): INTEGER: 3*
*            GGSN-MIB::ggsnPicIndex.2 = INTEGER: 2*
*            GGSN-MIB::ggsnPicAddress.2 = Wrong Type (should be IpAddress):
INTEGER: 4*
*            GGSN-MIB::ggsnPicNbrOfActivePdpContexts.2 = Wrong Type (should
be Gauge32 or Unsigned32): INTEGER: 6*
*            ........*
*            GGSN-MIB::ggsnPicIndex.50 = INTEGER: 50*
*            GGSN-MIB::ggsnPicAddress.50 = Wrong Type (should be IpAddress):
INTEGER: 100*
*            GGSN-MIB::ggsnPicNbrOfActivePdpContexts.50 = Wrong Type (should
be Gauge32 or Unsigned32): INTEGER: 150*

*        you can ignore the warning"Wrong Type (should be IpAddress)", and
just focus the output value which is correct.*
*
*
*        On the other hand the output of my printf in the function
example_table_get_next_data_point  is :
*
*

*
*

1,2,3,4,5,6,....199,200,1,2,3,....199,200,12,3,....,199,200.......1,2,3....199,200,
from
1 to 200, and repeated 50 times.
*
*

*
*To my understanding, it will cause performance problem when there are mass
of data. I have no idea that why we have to traverse *
*the whole 200 nodes every time to get a value, and in my opinion when we
use the snmptable command,  the operation to get the *
number (n) node should traverse the link-list the number (n-1) node, so I
think the output of the function *
example_table_get_next_data_point
*
*
when use the command snmptable should be as below :
*
*
           1,2,3,4,...........200
*
*
and the total iterations should be 200, but not 200*200, which will has a
strong impact on the efficiency.
*
*

*
Do I have any misunderstanding?  Look forward to your early reply and
some good advices for this problem.

Best Regards!

------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev

_______________________________________________
Net-snmp-users mailing list
[email protected]
Please see the following page to unsubscribe or change other options:
https://lists.sourceforge.net/lists/listinfo/net-snmp-users