[prev in list] [next in list] [prev in thread] [next in thread]
List: net-snmp-bugs
Subject: [ net-snmp-Bugs-409811 ] agentX mem leak if OID len > 10
From: noreply () sourceforge ! net
Date: 2001-03-30 17:32:03
[Download RAW message or body]
Bugs item #409811, was updated on 2001-03-19 11:46
You can respond by visiting:
http://sourceforge.net/tracker/?func=detail&atid=112694&aid=409811&group_id=12694
Category: agent
Group: None
Status: Open
>Priority: 3
Submitted By: Gopakumar H. Pillai (gopu)
Assigned to: John Naylon (jbpn)
Summary: agentX mem leak if OID len > 10
Initial Comment:
When an agentx sub-agent is running with an initial
OID length greater than 10 the agent leaks memory.
While using top to view the mem usage, I saw the agent
increasing in size by 8K chunks (probably top rounds
it to 8k) every few seconds.
How to reproduce:
I used the agentx sample to test it. I changed line 36
in ustScalarSet.c to simulate it.
original line:
static oid ustScalarSet_variables_oid[] = {
1,3,6,1,4,1,2021,13,4242,1,1 };
If compiled with the above line the program will leak.
edited line:
static oid ustScalarSet_variables_oid[] = {
1,3,6,1,4,1,1033,1,5 };
The above line would stop the leak.
Could any one help fixing this bug.
Thanks
--Gopu (gopu@hybrid.com)
----------------------------------------------------------------------
Comment By: John Naylon (jbpn)
Date: 2001-03-30 08:54
Message:
Logged In: YES
user_id=93926
Okay I think I have a patch which solves this problem
(attached). Please try it out and let us know if it works
for you.
John
----------------------------------------------------------------------
Comment By: John Naylon (jbpn)
Date: 2001-03-30 06:09
Message:
Logged In: YES
user_id=93926
This memory leak seems to occur when a getNext operation
leaves the registered range of the subagent. I am still
investigating.
----------------------------------------------------------------------
Comment By: John Naylon (jbpn)
Date: 2001-03-30 00:22
Message:
Logged In: YES
user_id=93926
Okay, in that case I *can* reproduce the bug. I don't have
a fix yet.
----------------------------------------------------------------------
Comment By: Gopakumar H. Pillai (gopu)
Date: 2001-03-29 10:31
Message:
Logged In: YES
user_id=148280
Hi JBPN,
Thanks for your efforts. I have been talking only about the
sub-agent leak and not at all the snmpd (master agent).
Master agent doesn't leak. But the sub-agent does, 8k
chunks every few minutes or seconds depending on the number
of queries receieved. Ofcourse the leak happens only if OID
len is greater than 10.
Hope this helps
--Gopu
----------------------------------------------------------------------
Comment By: John Naylon (jbpn)
Date: 2001-03-29 04:46
Message:
Logged In: YES
user_id=93926
I still cannot reproduce this bug. I see a small memory
leak in the *sub*agent, but nothing in the master agent.
Can you confirm that the leak you are talking about occurs
in the master agent?
----------------------------------------------------------------------
Comment By: John Naylon (jbpn)
Date: 2001-03-23 02:06
Message:
Logged In: YES
user_id=93926
Are there any other minor details you would like to
mention at this point?
----------------------------------------------------------------------
Comment By: Gopakumar H. Pillai (gopu)
Date: 2001-03-22 12:34
Message:
Logged In: YES
user_id=148280
I forgot to mention that, after running snmpd and ustMain
(the sample program), from another system I run snmpwalk in
an infinite loop. That's when the memory usage increases
--Gopu
----------------------------------------------------------------------
Comment By: Gopakumar H. Pillai (gopu)
Date: 2001-03-22 12:33
Message:
Logged In: YES
user_id=148280
I forgot to mention that, after running snmpd and ustMain
(the sample program), from another system I run snmpwalk in
an infinite loop. That's when the memory usage increases
--Gopu
----------------------------------------------------------------------
Comment By: Gopakumar H. Pillai (gopu)
Date: 2001-03-22 11:30
Message:
Logged In: YES
user_id=148280
I tried the CVS version and the memory leak still exists.
Code is same except that the init_master_agent call has
changed, now it does not take any arguments.
It takes about 2 to 3 minutes to see the leak of 8k on top
output
Hope this helps and thanks again
--Gopu
----------------------------------------------------------------------
Comment By: John Naylon (jbpn)
Date: 2001-03-22 09:29
Message:
Logged In: YES
user_id=93926
Could you try to reproduce the bug with the CVS version?
----------------------------------------------------------------------
Comment By: Gopakumar H. Pillai (gopu)
Date: 2001-03-22 09:24
Message:
Logged In: YES
user_id=148280
I am using UCD-SNMP version 4.2 .
I configured it with:
--with-mib-modules=agentx --prefix=/users/gopu/local7
I am attaching the whole config.h file (gzipped)
If u have more questions, u could send me a mail directly
(gopu@hybrid.com)
Thanks a million for taking these efforts.
--Gopu
----------------------------------------------------------------------
Comment By: John Naylon (jbpn)
Date: 2001-03-22 01:52
Message:
Logged In: YES
user_id=93926
Was this with the CVS version?
What did your configure line say?
----------------------------------------------------------------------
Comment By: Gopakumar H. Pillai (gopu)
Date: 2001-03-21 14:33
Message:
Logged In: YES
user_id=148280
I was trying on Solaris 7 on Sparc platform (Ultra 5 and
Ultra 10). I used top to see the size increasing, tested
for 2/3 days and the size was still increasing.
----------------------------------------------------------------------
Comment By: John Naylon (jbpn)
Date: 2001-03-21 02:06
Message:
Logged In: YES
user_id=93926
FWIW, I cannot reproduce this behaviour. Which platform is
this? Which version of the package? What does your
configure line say?
----------------------------------------------------------------------
Comment By: John Naylon (jbpn)
Date: 2001-03-20 09:45
Message:
Logged In: YES
user_id=93926
FWIW, I cannot reproduce this behaviour. Which platform is
this? Which version of the package? What does your
configure line say?
----------------------------------------------------------------------
You can respond by visiting:
http://sourceforge.net/tracker/?func=detail&atid=112694&aid=409811&group_id=12694
_______________________________________________
Net-snmp-bugs mailing list
Net-snmp-bugs@lists.sourceforge.net
http://lists.sourceforge.net/lists/listinfo/net-snmp-bugs
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic