Lots of "error on subcontainer 'ia_addr' insert (-1)" reports in /var/log/syslog

[Thomas Dreibholz]

Lots of "error on subcontainer ‘ia_addr’ insert (-1)" reports are written to /var/log/syslog. The problem *seems* to be IPv6 link-local addresses of GRE tunnels. Several GRE tunnels have the same IPv6 link-local address, which should be okay (since their scope is link-local). However, snmpd seems to have a problem with that.

ProblemType: Bug
DistroRelease: Ubuntu 12.04
Package: snmpd 5.4.3~dfsg-2.4ubuntu1.1
ProcVersionSignature: Ubuntu 3.2.0-55.85-generic 3.2.51
Uname: Linux 3.2.0-55-generic x86_64
ApportVersion: 2.0.1-0ubuntu17.6
Architecture: amd64
Date: Wed Oct 30 15:29:46 2013
InstallationMedia: Ubuntu-Server 12.04.1 LTS "Precise Pangolin" - Release amd64 (20120817.3)
MarkForUpload: True
SNMPVersion:
 NET-SNMP version: 5.4.3
 Web: http://www.net-snmp.org/
 Email: <email address hidden>
SourcePackage: net-snmp
SyslogSnmptrapd:

UpgradeStatus: No upgrade log present (probably fresh install)
modified.conffile..etc.snmp.snmpd.conf: [modified]
mtime.conffile..etc.snmp.snmpd.conf: 2013-10-11T12:02:23

[Thomas Dreibholz (dreibh)]

Modifying /etc/defaults/snmpd and setting "-Lsd" to "-LS6d" in SNMPDOPTS and TRAPDOPTS hides the error message (as suggested by http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=559109). At least, this should be the default then. However, somebody should have a closer look into the code in order to make sure that the link-local IPv6 addresses do not cause any problems.

[Thomas Dreibholz (dreibh)]

The Debian bug #559109 is quite old (from 2010). The current Ubuntu package (at least in 12.04 LTS) does *not* have this solution applied. It still uses the wrong logging level, producing the error output.

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Sina Sadeghi (sina-sa)]

This bug still appears to be present in Ubuntu 14.04

[Richard Laager (rlaager)]

I can confirm that `sed -i "s|-Lsd|-LS6d|" /etc/default/snmpd` fixes the problem for me.

[Nathan Neulinger (nneul-0)]

Still broken in 16.04 as well.

[Martend (martend-baas)]

Same "bug" here on 16.04.1 installation.

[Pik Master (pikmaster)]

Hi, the only option that I found, was to recompile the source with this patch.
I created a patch, to get rid of this error:

http://sourceforge.net/p/net-snmp/patches/1322/

Pik

[Christian Ehrhardt  (paelzer)]

This showed up "again" despite fixed and picked up in a merge due to switching to systemd.
The new Debian bug about it is https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=684721

[Kalle Tuulos (kalle-tuulos)]

Just reporting, that this bug still exists in 16.04.3 LTS.

[YG Kumar (ygk-kmr)]

Hi All,

I am using 'Ubuntu 16.04.2 LTS \n \l' . I see the following errors when starting snmpd:
----------
snmpd[130282]: error on subcontainer 'ia_addr' insert (-1)

Can someone help me to understand this error please ?

[Juan Carlos Ricardo (juanquy-i)]

Same error here.!
Any advances with it?

[Ivo Steinmann (isteinmann)]

Same error here with Ubuntu 17.04

[Mark Constable (markc)]

FWIW I'm still getting this error in my logs on 17.10 artful.

[Mark Constable (markc)]

Looks like I can't edit a comment. I found this works for me on 17.10...

sudo -i
cp /lib/systemd/system/snmpd.service /etc/systemd/system/snmpd.service
sed -i "s|-Lsd|-LS6d|" /etc/systemd/system/snmpd.service
systemctl daemon-reload
systemctl restart snmpd

[Andreas Hasenack (ahasenack)]

Confirmed in bionic:
Apr 05 17:44:58 bionic-snmpd snmpd[4740]: error on subcontainer 'ia_addr' insert (-1)
Apr 05 17:45:28 bionic-snmpd snmpd[4740]: error on subcontainer 'ia_addr' insert (-1)

[Andreas Hasenack (ahasenack)]

To clarify what I did:
- install snmpd
- install snmp-mibs-downloader
- add a nic alias. With ifconfig I did:
ifconfig eth0:1 10.0.100.209 netmask 255.255.255.0

- restart snmpd
- journalctl -u snmpd --follow

Wait a bit. The error should show up every 30s.

[carlos (nitrosito)]

Mee to, and not fix with sed -i "s|-Lsd|-LS6d|" /etc/default/snmpd

[N1nj4888 (n1nj4888)]

I'm still seeing this issue in Cosmic 18.10. It seems that, by default, snmpd is set to log at the DEBUG level which seems too intensive.

To fix I had to run the following (where "LS4d" indicates LOG_WARNING) as per the snmpd man page info below. With the log level set at LOG_WARNING this also has the added bonus that the many "Connection from UDP: [IP]:PORT->[IP]:PORT" snmpd INFO messages are also not output to /var/log/syslog...

$ sudo sed -i "s|-Lsd|-LS4d|" /lib/systemd/system/snmpd.service
$ sudo systemctl daemon-reload
$ sudo systemctl restart snmpd.service

snmpd man page (https://linux.die.net/man/1/snmpcmd):
0 or ! for LOG_EMERG,
1 or a for LOG_ALERT,
2 or c for LOG_CRIT,
3 or e for LOG_ERR,
4 or w for LOG_WARNING,
5 or n for LOG_NOTICE,
6 or i for LOG_INFO, and
7 or d for LOG_DEBUG.

[Christian Ehrhardt  (paelzer)]

Fixed in 5.8+dfsg-1 which we merged (we are at 5.8+dfsg-2ubuntu2)
So at lest Focal should be fixed by now.

[Christian Ehrhardt  (paelzer)]

Given how low prio/heat this issue had I'm gonna mark the SRU tasks won't fix.
Triaged+Low severity would be correct as well, but that would trigger wrong expectations as I just see no one get to it.

That way it is clear and if this really is more painful/severe than I thought people can speak up and outline why which then let us re-prioritize it.