No subject

=20
Request* req;
while ((run) && (!mib->get_agentx()->quit())) {
=20
req =3D reqList->receive(20000);
if (req) {
mib->process_request(req);
}
else {
// ping the master
mib->ping_master();
// we add or remove here rows=20
/* This shows how manipulation of a shared
table can be done from outside the table class:
if (ifTable->size() < 3) {
ifTable->any_index();
}
else {
List<MibTableRow>* l =3D ifTable->get_rows();
Oidx ind(l->last()->get_index());
l->clear();
delete l;
ifTable->remove_row(ind);
}
*/
}
}
=20
Bregards,
John Edmonds (BT)


________________________________

From: Frank Fock [mailto:fock at agentpp.com]
Sent: Mon 07/04/2008 16:02
To: Edmonds,J,John,JGFTX C
Cc: agentpp at agentpp.org
Subject: Re: [AGENT++] Core dump related to rows being removed from =
agentX tables?



Hello John,

With AGENT++ (and AgentX++) any access to a table's (or leaf's)
data must be synchronized. Some methods are already synchronized
which is documented in their method documentation.

Most methods are not synchronized. So those methods can be called
from within a synchronized method and outside such a method if
you are sure that this is safe.

Shortly spoken, to be thread-safe use:

mib->lock_mib();
table->start_synch();
mib->unlock_mib();
table->remove_row(index);
table->end_synch();

Best regards,
Frank Fock


john.edmonds at bt.com wrote:
> Hello Frank,
>=20
> Will you have any more information for us on this issue as this core =
dump problem is currently preventing us from deploying the code?
>=20
> Best regards,
> John Edmonds
>
> ________________________________
>
> From: Edmonds,J,John,JGFTX C
> Sent: Thu 03/04/2008 12:03
> To: Frank Fock
> Cc: agentpp at agentpp.org
> Subject: RE: [AGENT++] Core dump related to rows being removed from =
agentX tables?
>
>
> Hi Frank,
>=20
> No, I'm not explicitly calling the functions you mention.  I'm using =
the following, for example, to create tables and add and remove rows.   =
I based my code on the examples in the AgentX++ v1.4 documentation.
>=20
> ipsiRtpSenderEntry =3D new rtpSenderEntry("",mib);
> mib->add_no_reg(ipsiRtpSenderEntry);
>
> ipsiRtpSenderEntry->allocate_index(index)
>
> ipsiRtpSenderEntry->remove_row(index);
>
> From your e-mail it sounds like I might be missing something =
fundamental.   I guess I had assumed any mutex locking was done inside =
the above function calls.  Where/how do I make the calls you suggest?  =
Are there some documented examples I can follow?
>=20
> Best regards,
> John Edmonds (BT)
>=20
>
> ________________________________
>
> From: Frank Fock [mailto:fock at agentpp.com]
> Sent: Wed 02/04/2008 22:01
> To: Edmonds,J,John,JGFTX C
> Cc: agentpp at agentpp.org
> Subject: Re: [AGENT++] Core dump related to rows being removed from =
agentX tables?
>
>
>
> Hi John,
>
> How do you remove the rows from the table?
> Are you calling lock_mib() and table->start_synch()
> before removing the rows?
>
> Best regards,
> Frank
>
> john.edmonds at bt.com wrote:
>> Hi,
>>
>> I have a setup that works most of the time, but occasionally core =
dumps (under linux on PPC).  I suspect that it is somehow related to =
removing rows from tables (or multiple rows), when I don't remove table =
rows I never seem to get core dumps.  I am running an AgentX++ sub agent =
with Net-SNMP 5.4.1 master agent.
>>
>> agent++ v3.5.27.b
>> agentX++ v1.4.15
>> snmp++ v3.2.21.a
>>
>> This a stack backtrace I got from one core dump, the crash occurs at =
#0, List.h line 1154. Looks like a "get_next"...
>>
>> #7 thread_starter () at threads.cpp:486
>> #6 Agentpp::TaskManager::run () at threads.cpp:742
>> #5 Agentpp::MibTask::run () at threads.cpp:920
>> #4 Agentpp::SubAgentXMib::do_process_request () at =
agentx_subagent.cpp:926
>> #3 Agentpp::Mib::process_request () at mib.cpp:3297
>> #2 Agentpp::SubAgentXMib::find_next () at agentx_subagent.cpp:1072
>> #1 Agentpp::MibTable::find_succ () at mib.cpp:1753
>> #0 Agentpp::MibTable::find_next () at List.h:1154
>>
>> I have a MIB Browser that is constantly refreshing the display of the =
Tables.  Could there be a missing mutex somewhere between table row =
updates and browser SNMP gets perhaps?  Or confusion between a get on =
one table and row removal on another?
>>
>> Any suggestions as to what might be happening or what I could do to =
patch code to stop core dumps?  Add some bounds checking?
>>
>> Best regards,
>> John Edmonds (BT)
>>
>>
>>
>>
>>
>>
>>
>> _______________________________________________
>> AGENTPP mailing list
>> AGENTPP at agentpp.org
>> http://lists.agentpp.org/mailman/listinfo/agentpp
>
> --
> AGENT++
> http://www.agentpp.com <http://www.agentpp.com/>  =
<http://www.agentpp.com/>
> http://www.mibexplorer.com <http://www.mibexplorer.com/>  =
<http://www.mibexplorer.com/>
> http://www.mibdesigner.com <http://www.mibdesigner.com/>  =
<http://www.mibdesigner.com/>
>
>
> _______________________________________________
> AGENTPP mailing list
> AGENTPP at agentpp.org
> http://lists.agentpp.org/mailman/listinfo/agentpp

--
AGENT++
http://www.agentpp.com <http://www.agentpp.com/>=20
http://www.mibexplorer.com <http://www.mibexplorer.com/>=20
http://www.mibdesigner.com <http://www.mibdesigner.com/>=20