Hide Forgot
Description of problem: We have a RHEL6 x86_64 server onto which we've installed RT, and imported several hundred gigs of email for an audit. We have backed up and restored the RT database a couple times to ensure that backups and restores are successful. As such, we've written and read an enormous amount of data to the SAN volume which holds the database. We use Zenoss to monitor this server, and during every polling cycle we get the following error in /var/log/messages: Mar 3 11:49:07 servername snmpd[17720]: truncating integer value > 32 bits Using snmpwalk we have identified that the error occurs only when querying the following OIDs: 1.3.6.1.4.1.2021.11.57.0 1.3.6.1.4.1.2021.11.58.0 Version-Release number of selected component (if applicable): net-snmp-libs-5.5-27.el6_0.1.x86_64 net-snmp-5.5-27.el6_0.1.x86_64 net-snmp-utils-5.5-27.el6_0.1.x86_64 How reproducible: Fully reproducible. Every time an SNMP query hits either ssIORawSent or ssIORawReceived the error will be logged. Steps to Reproduce: 1. snmpwalk -v 1 -c foobar 1.2.3.4 1.3.6.1.4.1.2021.11.57.0 2. snmpwalk -v 1 -c foobar 1.2.3.4 1.3.6.1.4.1.2021.11.58.0 3. repeat the above two commands with -v 2c, too Actual results: snmpd[17720]: truncating integer value > 32 bits Expected results: no error Additional info: It looks like this may be related to the following two closed bugs: https://bugzilla.redhat.com/show_bug.cgi?id=202584 https://bugzilla.redhat.com/show_bug.cgi?id=528164 I suspect that a reboot of the server will clear up these errors, and we're not likely to generate this much disk I/O again any time soon, so this is not a critical error; but it seems like the kind of thing that net-snmp should account for properly.
I fixed it upstream: http://net-snmp.svn.sourceforge.net/viewvc/net-snmp?view=revision&revision=20061
Technical note added. If any revisions are required, please edit the "Technical Notes" field accordingly. All revisions will be proofread by the Engineering Content Services team. New Contents: The previous version of snmpd, the Net-SNMP agent, gathered disk IO and CPU usage statistics for UCD-SNMP::systemStats as 64bit on 64bit systems. Appropriate MIB describes these statistics as 32bit and snmpd reported "truncating integer value > 32 bits" to system log when processing the 64bit values. snmpd in updated packages collects values of UCD-SNMP::systemStats as 32bit integers and thus does not report aforementioned message to syslog. (BZ#681949).
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. http://rhn.redhat.com/errata/RHBA-2011-1524.html