Hi Robert,
Thanks for the reply.
Please find below the valgrind report for the growing leak :
a) Below report when snmp is up and running :

255 (108 direct, 147 indirect) bytes in 3 blocks are definitely lost in 
loss record 105 of 153
==20680== at 0x4004864: calloc (vg_replace_malloc.c:279)
==20680== by 0x83E761C: netsnmp_access_ipaddress_entry_create 
(ipaddress_common.c:175)
==20680== by 0x84BEEFA: _netsnmp_ioctl_ipaddress_container_load_v4 
(ipaddress_ioctl.c:174)
==20680== by 0x84BE444: netsnmp_arch_ipaddress_container_load 
(ipaddress_linux.c:149)
==20680== by 0x83E750A: netsnmp_access_ipaddress_container_load 
(ipaddress_common.c:120)
==20680== by 0x84BDA70: ipAddressTable_container_load 
(ipAddressTable_data_access.c:345)
==20680== by 0x84BD370: _cache_load (ipAddressTable_interface.c:1922)
==20680== by 0x84CC524: _cache_load (cache_handler.c:548)
==20680== by 0x84CBDED: _timer_reload (cache_handler.c:197)
==20680== by 0x84E86D2: run_alarms (snmp_alarm.c:266)
==20680== by 0x83BFF7F: receive (snmpd.c:1504)
==20680== by 0x83BF3A4: snmpd (snmpd.c:1177)

b) Below report when i left snmp running for more than one hour:

10,200 (4,320 direct, 5,880 indirect) bytes in 120 blocks are definitely 
lost in loss record 136 of 154
==20742== at 0x4004864: calloc (vg_replace_malloc.c:279)
==20742== by 0x83E761C: netsnmp_access_ipaddress_entry_create 
(ipaddress_common.c:175)
==20742== by 0x84BEEFA: _netsnmp_ioctl_ipaddress_container_load_v4 
(ipaddress_ioctl.c:174)
==20742== by 0x84BE444: netsnmp_arch_ipaddress_container_load 
(ipaddress_linux.c:149)
==20742== by 0x83E750A: netsnmp_access_ipaddress_container_load 
(ipaddress_common.c:120)
==20742== by 0x84BDA70: ipAddressTable_container_load 
(ipAddressTable_data_access.c:345)
==20742== by 0x84BD370: _cache_load (ipAddressTable_interface.c:1922)
==20742== by 0x84CC524: _cache_load (cache_handler.c:548)
==20742== by 0x84CBDED: _timer_reload (cache_handler.c:197)
==20742== by 0x84E86D2: run_alarms (snmp_alarm.c:266)
==20742== by 0x83BFF7F: receive (snmpd.c:1504)
==20742== by 0x83BF3A4: snmpd (snmpd.c:1177)


You can see that Leak ( definitely lost memory ) is increased from 255 
to 10,200 bytes.
Why the run_alarms function calls the function to load the container 
always if it already has been done once.
I am new to the container concept, please give me some pointers to work 
with this issue.

Regards,
Manjit



Robert Story wrote:
> On Fri, 07 May 2010 17:36:18 +0530 Manjit wrote:
> M> ==698==    by 0x84BD990: ipAddressTable_container_load 
> M> Please provide me information for the below queries :
> M> 1. what is the purpose of run_alarms() ?
>
> Just what is says... It runs alarms, or events.
>
> M> 2. How the system will be affected if i will not call (or  comment ) 
> M> function ?
>
> It's possible that lots of things depending on alarms to be run at regular
> intervals will break.  The problem is not with run_alarms itself, but with the
> particular alarm being run.  Try running with -I-ipAddressTable to disable the
> ipAddressTableand it's alarm, though I suspect it's being called via ifTable.
>
> I'm also not sure memory is being lost. Are you actually seeing the memory
> leak growing? Lots of places in net-snmp allocate memory that isn't explicitly
> released when shutting down, but they don't cause growing leaks during 
> runtime.
>
>   


------------------------------------------------------------------------------

_______________________________________________
Net-snmp-coders mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/net-snmp-coders

Reply via email to