Hi,
I have strange error messages logged when I have bonding interface.
OS: FC5
net-snmp-5.3-4.2
Kernel: 2.6.19-1.2288.2.1.fc5 #1 SMP Thu Feb 22 11:32:04 EST 2007 x86_64
This is my configuration file:
############### Generic config information ###############
syslocation "F4, UK"
sysservices 0
syscontact Company NOC noc@company.com
rocommunity r3ad0n1y
trapcommunity r3ad0n1y
trapsink 1.2.3.4
############### Common config directives ###############
disk /
disk /var
disk /usr
disk /tmp
swap 16000
linkUpDownNotifications yes
defaultMonitors yes
master agentx
############### Node specific config directives ###############
proc asterisk 0 0
I have Bonding interface and my two physical interfaces are configured as follow:
DEVICE=(eth0 and eth1 for second interface)
USERCTL=no
ONBOOT=yes
MASTER=bond0
SLAVE=yes
BOOTPROTO=none
ETHTOOL_OPTS="speed 100 duplex full autoneg off"
Now every 20 seconds I have in my logs:
Apr 2 16:02:48 f-node0 snmpd[9588]: error on subcontainer 'ia_addr' insert (-1)
Apr 2 16:02:48 f-node0 snmpd[9588]: error on subcontainer 'ia_addr' insert (-1)
Apr 2 16:02:48 f-node0 snmpd[9588]: error on subcontainer '' insert (-1)
Apr 2 16:02:48 f-node0 snmpd[9588]: error on subcontainer '' insert (-1)
Apr 2 16:02:48 f-node0 snmpd[9588]: failed to run mteTrigger query
Apr 2 16:02:48 f-node0 snmpd[9588]: failed to run mteTrigger query
I suppose that the bonding slaves are the problems as when I stop for example eth1 I have two messages less:
error on subcontainer 'ia_addr' insert (-1)
error on subcontainer '' insert (-1)
BR,
ddragoev@gmail.com
Logged In: YES
user_id=1552812
Originator: NO
You have some Interfaces with the same Interface-Address.
The Index if the IP-Address-Table is only the Address (Use unique IF-IP-Addresses).
The error happends every 30sec (refresh timer).
The "linkUpDownNotifications yes" - command (DISMAN-MIB) triggers the Table-Refresh
The command "ifconfig" displays only the primary address.
Check with "ip addr show" (iproute2-Package)
Bye,
Alex
Logged In: YES
user_id=1230171
Originator: NO
Hi Alex,
Thanks’ for your reply. However I don’t have same IP address on 2 or more interfaces. May be the problem is that I have same MAC address but it is set from bonding driver and I can’t do anything. The other strange is if I remove linkUpDownNotifications the errors continue.
1: lo: <LOOPBACK,UP,10000> mtu 16436 qdisc noqueue
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: eth0: <BROADCAST,MULTICAST,SLAVE,UP,10000> mtu 1500 qdisc pfifo_fast master bond0 qlen 1000
link/ether 00:11:25:c4:52:d4 brd ff:ff:ff:ff:ff:ff
inet6 fe80::211:25ff:fec4:52d4/64 scope link
valid_lft forever preferred_lft forever
3: eth1: <BROADCAST,MULTICAST,SLAVE,UP,10000> mtu 1500 qdisc pfifo_fast master bond0 qlen 1000
link/ether 00:11:25:c4:52:d4 brd ff:ff:ff:ff:ff:ff
inet6 fe80::211:25ff:fec4:52d4/64 scope link
valid_lft forever preferred_lft forever
4: bond0: <BROADCAST,MULTICAST,MASTER,UP,10000> mtu 1500 qdisc noqueue
link/ether 00:11:25:c4:52:d4 brd ff:ff:ff:ff:ff:ff
inet 192.168.0.177/27 brd 192.168.0.191 scope global bond0
inet6 fe80::211:25ff:fec4:52d4/64 scope link
valid_lft forever preferred_lft forever
Logged In: YES
user_id=1782758
Originator: NO
This error actually causes a memory leak, as the container is not cleaned up when the "insert" fails. I have a patch, and will submit it when I figure out how :)
Logged In: YES
user_id=1782758
Originator: NO
Submitted patch 1710632 (as a diff).
This patch can be applied to net-snmp 5.4 aswell (even though I was working with 5.3)
Logged In: YES
user_id=1784238
Originator: NO
Submitted patch #1712645, which prints more meaningful error message and prints it only once
Logged In: YES
user_id=588340
Originator: NO
Its not a problem with bonding - its something to do with duplicate IP addresses (which encompasses bonding and tunnels etc)
Logged In: NO
This issue only seems to occur when the bonding driver is in play. Furthermore, I used the patch for 5.3.1 against the net-snmp-5.3.1-14.0.1.el5.src.rpm src rpm for CentOS 5 and snmpd segfaults on startup if eth1 is the active interface and not eth0
Logged In: NO
HI, i have it problem to 8(( whith 'veth' interface
Linux 2.6.21-xen (Gentoo)
eth0,veth0 ......
What can i do ??
Logged In: YES
user_id=848638
Originator: NO
The problem should be fixed in SVN Revision 16804 on 2008-02-14.
Can you please try with 5.4.2.pre1 (or later) or a recent SVN snapshot, either build from source
(http://www.net-snmp.org/nightly/tarballs) or a binary SVN snapshot package
(http://www.net-snmp.org/wiki/index.php/Third-Party_Packages) and report
back?
Logged In: NO
Thomas,
I just recently built a 5.4.1.2 agent using the 5.4.1.2 code I downloaded from the site. On an AS3 system with bonded interfaces I got the following:
NET-SNMP version 5.4.1.2
error on subcontainer 'ia_addr' insert (-1)
error on subcontainer 'ia_index' insert (-1)
error on subcontainer 'ia_index' insert (-1)
error on subcontainer 'ia_addr' insert (-1)
error on subcontainer '' insert (-1)
The agent proceeded to stop responding to requests. I'm doing more testing to confirm on a different system.
Logged In: YES
user_id=848638
Originator: NO
We already know that the problem is *not* fixed in 5.4.1.2 -- which is exactly the reason why I've been asking for verification with 5.4.2.rc1 or later where it should be fixed. 5.4.2.rc2 is out so please try that one and report back.
Thanks for the bug report!
We've fixed the problem in the 5.3.x and 5.4.x
code branches and the main development tree,
so it should be fixed in future releases of the
Net-SNMP package.
Appropriate log messages shall now have lower severity (LOG_DEBUG), so you should not see them with default settings. SVN rev. 17254.
backported also to 5.2.x, SVN rev. 17257
I am running 5.4.2.1 built on Solaris SPARC.
compiled with --with-mib-modules=if-mib --enable-mfd-rewrites; runs fine on global zone but in the NON-global zone, the process is running but can't query and get the following error in the /var/log/snmpd.log
error on subcontainer 'interface container' insert (-1)
I even tried applying the patch but no prevail.
'interface container' is different container, the patch won't help you here. All the 'error on subcontainer' are mostly harmless and they will have lower log level in future releases on net-snmp (e.g. 5.4.3 and 5.5).
Hi Thomas and Jan,
I find the similar error in NET-SNMP version: 5.5
I find the following error in snmpd.log:
mibII/mta_sendmail.c:open_sendmailst: could not guess version of statistics file "/var/log/mail/statistics"
error on subcontainer 'ia_addr' insert (-1)
NET-SNMP version 5.5
error on subcontainer 'ia_addr' insert (-1)
Is your fix included in version 5.5?
Yes, it is included in net-snmp 5.5.2. You may still see the message if you have enabled 'debug' level messages in snmpd.log.
Hi Jan
I'm getting below messages updated whenever snmp restarts.
root@sandy01> cat /var/log/snmpd.log
error on subcontainer 'ia_addr' insert (-1)
Duplicate IPv4 address detected, some interfaces may not be visible in IP-MIB
error on subcontainer 'ia_addr' insert (-1)
error on subcontainer 'ia_addr' insert (-1)
error on subcontainer 'ia_addr' insert (-1)
error on subcontainer 'ia_addr' insert (-1)
error on subcontainer 'ia_addr' insert (-1)
Turning on AgentX master support.
NET-SNMP version 5.5
root@sandy01>
1.what is the cause for "Duplicate Ipv4 address " ?
2.How to avoid it?
kernel 2.6.32-431.23.3.el6.x86_64
net-snmp-utils-5.5-49.el6_5.1.x86_64
net-snmp-libs-5.5-49.el6_5.1.i686
net-snmp-libs-5.5-49.el6_5.1.x86_64
net-snmp-5.5-49.el6_5.1.x86_64
net-snmp-perl-5.5-49.el6_5.1.x86_64
Is this fixed already?
Best Regards
Rajkumar
"Duplicate IPv4 address detected, some interfaces may not be visible in IP-MIB" means exactly what it says. There are two or more network interfaces with the same IP address. Since the IP address is key column in several IP-MIB tables, only one the the interfaces will appear there. It is not a bug in Net-SNMP.