Opened 11 years ago

Last modified 9 years ago

#65 assigned defect

Ubuntu 10.4: SMTP daemon dies after 24 hrs

Reported by: suehring@… Owned by: bas
Priority: major Milestone:
Version: 6.0.1 Keywords: dsm_sa_snmp32d, segfault, crash, ubuntu, 10.4
Cc:

Description

I have installed the new Ubuntu release (64-bit version) on a Dell PowerEdge? 1950 server. After adding libstdc++5 and the 32-bit PAM libraries everything (including web interface and SNMP daemon) seems to start properly.

Only after exactly one day (24 hours) the SNMP daemon dies (/var/log/messages):

dsm_sa_snmp32d[30195]: segfault at 0 ip 00000000f70b7565 sp 00000000f7062ba0 error 4 in libdsm_sm_ssimp32.so.3.0.2.553[f70a7000+15000]

I suspect that a log rotation function might be trying to access a file without the proper permission. But I could not find any other log entry that is helpful for debugging.

Do you have any ideas what might be causing this?

Change History (9)

comment:1 Changed 11 years ago by bas

  • Owner set to bas
  • Status changed from new to assigned

Just a guess we do no use snmp daemon. Do you do some cleaning of /tmp or /var/tmp. Maybe it uses a socket file in one of these directories.

comment:2 Changed 11 years ago by anonymous

lsof does not show any open files in /tmp or /var/tmp.

I don't think it's an cron based cleanup process because the process stops 24 hours after it has been started, not at a specific time.

comment:3 Changed 11 years ago by anonymous

Actually i have the same problem, but it seems more like a memory hole. The process grows until it reaches ~2GB RAM, then it segfaults. My workaround is to restart dataeng every morning...

comment:4 Changed 11 years ago by anonymous

You are right. top shows 2.9 GB after approx 20 hours.

The same process takes only 1.5 MB on Ubuntu Jaunty machine after 25 days uptime.

comment:5 Changed 11 years ago by suehring@…

I'm using the check_openmanage NAGIOS plugin. Each call seems to add ~15 MB of memory usage to dsm_sa_snmp32d.

A simple

snmpwalk -v 1 -c public localhost .1.3.6.1.4.1.674.10892.1

does not affect the memory usage although the Dell specific entries are displayed.

comment:6 Changed 11 years ago by sstdennis@…

Ditto (Ubuntu 10.4; Dell OMSA 6; Nagios check_openmanage plugin). Some machine have dsm_sa_snmp32d die daily, others once or twice a week.

I have put in cron entries to restart OMSA twice a day. I will see if this fixes the symptoms. I await a solution to the root problem.

comment:7 Changed 10 years ago by anonymous

I experience the same problem - did anyone find a solution ? thanks

comment:8 Changed 9 years ago by adrian@…

Still happens (calling with -B or -e -B). Packages: ii libsnmp-base 5.4.3~dfsg-2.1 SNMP (Simple Network Management Protocol) MIBs and documentation ii libsnmp15 5.4.3~dfsg-2.1 SNMP (Simple Network Management Protocol) library ii snmpd 5.4.3~dfsg-2.1 SNMP (Simple Network Management Protocol) agents ii dellomsa 6.0.1-10

comment:9 Changed 9 years ago by adrian@…

Ooh - 6.4 apparently fixes this.

(dropped http prefix off to try and bypass spam filter (which seems to try and present a captcha, but fails to do so) en.community.dell.com/techcenter/b/techcenter/archive/2011/01/13/dell-openmanage-server-administrator-6-4-for-ubuntu.aspx

"fix for a memory leak issue with dsm_sa_snmpd that occurs when you use the Nagios check_openmanage plugin"

Note: See TracTickets for help on using tickets.