Re: how to specify source IP address / interface when sending traps (with Solaris IP multipathing )

Tue, 16 Aug 2005 16:51:44 -0700 


Okay here's the final word on this thread:
The change of importance was in net-snmp/snmplib/snmpUDPDomain.c between revision 5.10 and 5.11. It happened 23 months ago. I don't really following the naming convention used on CVS branches in the tree, but it was pulled into the 5.1 branch obviously before 5.1.3. 

http://cvs.sourceforge.net/viewcvs.py/net-snmp/net-snmp/snmplib/snmpUDPDomain.c?r1=5.10&r2=5.11

Whoever commit the pullup forgot the sister change to man/snmp.conf.5.def
adding "clientaddr [ip]" into your snmp.conf (NOTE: NOT SNMPD.CONF) will change how the bind() works in 5.1.3. 

Welcome to UNIX daemons in the 21st century.
I will try to work with the OpenBSD people to get the port updated 5.1.3, but it doesn't look very promising. Honestly, all of this work is the result of the OpenBSD port being behind (probably due to the patches). When it's all said and done, I wonder if it wouldn't just make more sence ot get 5.2.x running instead. This would be a non-issue if the port was up-to-date in the 5.1 branch or was in the 5.2 branch. A lot of the OpenBSD ports patches to 5.1.2 have apparently been merged back into the main tree in 5.1.3.
Could someone please pull up the snmp.conf.5 into the 5.1 branch, as well as possibly modify the text: 

.IP "clientaddr [<transport-specifier>:]<transport-address>"
Sender address used by the command-line tools, see
...it should really say, "Specify source address used by command-line tools and the agent daemon snmpd(8) when sending traps and informs....etc." 


===>  Patching for net-snmp-5.1.3p2
1 out of 3 hunks failed--saving rejects to agent/mibgroup/mibII/tcpTable.c.rej 
***>   patch-agent_mibgroup_mibII_tcpTable_c did not apply cleanly
1 out of 1 hunks failed--saving rejects to agent/mibgroup/ucd_snmp.h.rej
***>   patch-agent_mibgroup_ucd_snmp_h did not apply cleanly
1 out of 7 hunks failed--saving rejects to configure.rej
***>   patch-configure did not apply cleanly
Ignoring previously applied (or reversed) patch.
4 out of 4 hunks ignored--saving rejects to local/fixproc.rej
***>   patch-local_fixproc did not apply cleanly
*** Error code 1


~BAS




On Tue, 16 Aug 2005, Brian A. Seklecki wrote:


On Tue, 16 Aug 2005, Brian A. Seklecki wrote:


On Tue, 16 Aug 2005, Russell Bell wrote:



Is the problem definitely fixed in 5.1.3 then ?  Re the code similarity,
Matt copied the WIN32 version of get_myaddr() in 5.1.2 snmplib/system.c and
I'm going to find out this morning. The changelog between 5.1.2 and 5.1.3 is something like 30 pages long. None of the OpenBSD ports patches apply cleanly. 

Russell:

Could you please send me your copy of system.c?
I was able to safely backport all of _sess_open() from snmplib/snmp_api.c from 5.2.1.2 from to 5.1.3 by simply copying it and also a small chunk from the top of snmp_api.c: 

   netsnmp_ds_register_config(ASN_OCTET_STR, "snmp", "clientaddr",
                     NETSNMP_DS_LIBRARY_ID, NETSNMP_DS_LIB_CLIENT_ADDR);
These are the only places that I see 'clientaddr' mentioned...but I don't see a mention of the config paramter until snmp.5.conf in -current sources. It occured to me that 'clientaddr' is read from snmp.conf instead of snmpd.conf.
Anyway, my snmpd(8) agent daemon still isn't properly sourcing traps with that patch applied.
Do config file statements in snmp.conf apply to all functions/code, even if traps are sent from snmpd(8) through shared code ...instead of snmptrap(1) ? 

~BAS


-------------------------------------------------------
SF.Net email is Sponsored by the Better Software Conference & EXPO
September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
_______________________________________________
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



l8*
        -lava

x.25 - minix - bitnet - plan9 - 110 bps - ASR 33 - base8


-------------------------------------------------------
SF.Net email is Sponsored by the Better Software Conference & EXPO
September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
_______________________________________________
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

Reply via email to