All,
I have tried to get to the bottom of this problem -
but this seems to be bottomless pit for me!!
I have found the following:
- If I did a snmpbulkget with just one oid
specified, even if it spans the mib tree on 2
different subagents, they seem to work fine! But the
problem happens whenever I try to do the same with 2
or more oids - and all of them should fetch oids in
trees supported by other subagents - then I have
SIGSEGV.
- Here's the stacktrace, when all works fine (i.e.,
with 1 oid in snmpbulkget):
#0 netsnmp_table_iterator_helper_handler
(handler=0x80d0a90,
reginfo=0x80caf10, reqinfo=0x80e55d0,
requests=0x80e0390)
at table_iterator.c:435
#1 0x2832e510 in netsnmp_call_next_handler
(current=0x283ce4f8,
reginfo=0x80caf10, reqinfo=0x80e55d0,
requests=0x80e0390)
at agent_handler.c:423
#2 0x283d2a82 in table_helper_handler
(handler=0x80d0b10, reginfo=0x80caf10,
reqinfo=0x80e55d0, requests=0x80e0390) at
table.c:614
#3 0x2832e0eb in netsnmp_call_handlers
(reginfo=0x80caf10, reqinfo=0x80e55d0,
requests=0x80e0390) at agent_handler.c:423
#4 0x283241f0 in handle_var_requests (asp=0x80e0310)
at snmp_agent.c:2394
#5 0x28324c34 in handle_pdu (asp=0x80e0310) at
snmp_agent.c:3171
#6 0x28325b2c in netsnmp_handle_request
(asp=0x80e0310, status=135148176)
at snmp_agent.c:2967
#7 0x2832601d in handle_snmp_packet (op=1,
session=0x0, reqid=1307, pdu=0x0,
magic=0x80e6010) at snmp_agent.c:1792
#8 0x2840a33a in _sess_process_packet
(sessp=0x80cee10, sp=0x80cd210,
isp=0x80ca790, transport=0x80ca290,
opaque=0x80e3290, olength=4,
packetptr=0x80fd010 "", length=1) at
snmp_api.c:5208
#9 0x2840b450 in _sess_read (sessp=0x80cee10,
fdset=0xbfbfe060)
at snmp_api.c:5606
#10 0x2840bdc0 in snmp_sess_read (sessp=0x80cee10,
fdset=0xbfbfe060)
at snmp_api.c:5625
#11 0x2840be16 in snmp_read (fdset=0xbfbfe060) at
snmp_api.c:5260
#12 0x284691e6 in snmpSubAgentReadCallBackFunc ()
from
/home/arijit/beta2-mainline-branch/pkg/olive/ansr/lib/libsnmpsubagent.so
- Here's the stack trace when we get the SIGSEGV:
#0 0x283cec35 in
netsnmp_table_iterator_helper_handler
(handler=0x80d0a90,
reginfo=0x80caf10, reqinfo=0x80e5f50,
requests=0x80e1910)
at table_iterator.c:435
#1 0x2832e510 in netsnmp_call_next_handler
(current=0x283ce4f8,
reginfo=0x80caf10, reqinfo=0x80e5f50,
requests=0x80e1910)
at agent_handler.c:423
#2 0x283d2a82 in table_helper_handler
(handler=0x80d0b10, reginfo=0x80caf10,
reqinfo=0x80e5f50, requests=0x80e1910) at
table.c:614
#3 0x2832e0eb in netsnmp_call_handlers
(reginfo=0x80caf10, reqinfo=0x80e5f50,
requests=0x80e1910) at agent_handler.c:423
#4 0x283241f0 in handle_var_requests (asp=0x80e0310)
at snmp_agent.c:2394
#5 0x28324c34 in handle_pdu (asp=0x80e0310) at
snmp_agent.c:3171
#6 0x28325b2c in netsnmp_handle_request
(asp=0x80e0310, status=135148144)
at snmp_agent.c:2967
#7 0x2832601d in handle_snmp_packet (op=1,
session=0x0, reqid=1182, pdu=0x0,
magic=0x0) at snmp_agent.c:1792
#8 0x2840a33a in _sess_process_packet
(sessp=0x80cee10, sp=0x80cd210,
isp=0x80ca790, transport=0x80ca290,
opaque=0x80e3270, olength=4,
packetptr=0x80f9010 "", length=1) at
snmp_api.c:5208
#9 0x2840b450 in _sess_read (sessp=0x80cee10,
fdset=0xbfbfe060)
at snmp_api.c:5606
#10 0x2840bdc0 in snmp_sess_read (sessp=0x80cee10,
fdset=0xbfbfe060)
at snmp_api.c:5625
#11 0x2840be16 in snmp_read (fdset=0xbfbfe060) at
snmp_api.c:5260
#12 0x284691e6 in snmpSubAgentReadCallBackFunc ()
from
/home/arijit/beta2-mainline-branch/pkg/olive/ansr/lib/libsnmpsubagent.so
From these stacktraces - one glaring difference is
that the magic passed over from
_session_process_packet to netsnmp_handle_request is
null(0x0), when I get the SIGSEGV, but not in the case
where all works fine.
I am not postive thats the problem here, but I have
noted that "magic" is the root of information for
"requests" whose parent_data is null which causes the
ultimate segmentation violation.
If thats the case, does that mean I have some slp
node in the Sessions linkedlist that gets closed when
first time the snmpbulkget spans multiple subagents,
so, the next time the call tries to reach the other
subagent using the same session - it gets screwed?
I am most probably halucinating here, but thought,
I would mention my mental state in trying to get to
the bottom of this issue - hoping someone more
literate takes pity and points me to the root cause :)
Eagerly looking forward to some pointer (light at
the bottom of this endless pit)
-Arijit
--- arijit <[EMAIL PROTECTED]> wrote:
> All,
> I have a few subagents which support their MIB
> trees
> and talk to the main agent using agentx. Under
> normal
> walk - all works fine. Even when I do a bulkget with
> one oid specified - I easily get the next node
> supported by another subagent on the tree.
> However, the fun happens when I try to do bulkget
> with more than 1 oid specified (I am using 0
> non-repeaters). So, as soon as the bulkget causes
> the
> entire tree supported on this subagent to get
> exhausted, the subagent crashes with signal 11.
> I havent yet done a detailed debugging of the
> cause
> - but was just wondering if anyone already faced
> this
> problem!
> Thanks and regards,
> -Arijit
>
> __________________________________________________
> Do You Yahoo!?
> Tired of spam? Yahoo! Mail has the best spam
> protection around
> http://mail.yahoo.com
>
>
>
-------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support
> web services, security?
> Get stuff done quickly with pre-integrated
> technology to make your job easier
> Download IBM WebSphere Application Server v.1.0.1
> based on Apache Geronimo
>
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> 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
>
__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
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
