[AGENT++] SNMP++ Potential Memory Leak
Stoyan Haralampiev
tony at gyuvetch.bg
Wed Nov 1 08:50:21 CET 2006
I am posting this message, so that its clear on the Mailing List
Archive which thread this solution belongs to.
Here's the original thread that the solution has to do with:
http://lists.agentpp.org/pipermail/agentpp/2006-October/003729.html
>-------- Оригинално писмо --------
>От: Stoyan Haralampiev
>Относно: Re: Re: [AGENT++] SNMP++ Potential Memory Leak
>До: Jochen Katz
>Изпратено на: Сряда, 2006, Ноември 1 09:43:18 GMT+02:00
>----------------------------------
>
>
>Hi, Jochen,
>
>Bravo! Great you found it!
>
>Thank you very much for the quick and effective support!
>
>Best regards,
>Tony
>
>
>
>>-------- Оригинално писмо --------
>>От: Jochen Katz
>>Относно: Re: [AGENT++] SNMP++ Potential Memory Leak
>>До: Stoyan Haralampiev
>>Изпратено на: Вторник, 2006, Октомври 31 23:14:31 GMT+02:00
>>----------------------------------
>>
>>Hi Tony,
>>
>>I think I found it.
>>
>>> perl script that forks "snmptrap" processes as fast as it can. So the
>>> trap sender sends packets at a very high rate.
>>
>>This leads to the problem, that each snmptrap process will use a
>>different source port and receive_trap adds the combination engine
>>id/host/port into its tables. The current version checks only for
>>duplicate host/port. The attached patch will also check for duplicate
>>engine ids.
>>
>>This patch will go into the next snmp++ release.
>>
>>Regards,
>> Jochen
>>
>>
>>
>
>-----------------------------------------------------------------
>Survivor BG. Оцеляването продължава на www.survivor.btv.bg
>_______________________________________________
>AGENTPP mailing list
>AGENTPP at agentpp.org
>http://lists.agentpp.org/mailman/listinfo/agentpp
>
-----------------------------------------------------------------
Survivor BG. Оцеляването продължава на www.survivor.btv.bg
More information about the AGENTPP
mailing list