Just now I found out that the /var partition on our Solaris hosting SBR Carrier 7.2.4 has reached 100% capacity due to large core files located in /var/core/ directory in the format of core_<hostname>_jnprsnmpd_<numbers>.
root@SBRKO03-RADFE02 # ls -lthr total 56712224 -rw------- 1 root root 4.0G May 16 16:33 core_SBRKO03-RADFE02_jnprsnmpd_0_0_1337160612_25009 -rw------- 1 root root 4.0G Jun 1 04:00 core_SBRKO03-RADFE02_jnprsnmpd_0_0_1338497846_17861 -rw------- 1 root root 4.0G Jun 17 23:30 core_SBRKO03-RADFE02_jnprsnmpd_0_0_1339950465_28722 -rw------- 1 root root 4.0G Jul 13 06:18 core_SBRKO03-RADFE02_jnprsnmpd_0_0_1342134962_20197 -rw------- 1 root root 4.0G Aug 2 15:23 core_SBRKO03-RADFE02_jnprsnmpd_0_0_1343895621_29822 -rw------- 1 root root 4.0G Aug 18 21:30 core_SBRKO03-RADFE02_jnprsnmpd_0_0_1345300058_1566 -rw------- 1 root root 3.1G Sep 3 19:00 core_SBRKO03-RADFE02_jnprsnmpd_0_0_1346673474_13203 root@SBRKO03-RADFE02 #
I suspected that these are core dumps from jnprsnmpd process, but I was hoping you guys can give more insightful information on them. What triggers the creation of these jnprsnmpd core files? Is it safe to remove it manually? Can we configure the jnprsnmpd process to automatically delete old core files?
Looking forward for your answers.
Solved! Go to Solution.
General recommendation is to install SBRC in /opt partition. We need to look in to the SBR debug log file and SNMP log taken during this coring to see what is causing the issue however to know exact cause of this issue the core should be reviewed.
For this you need to contact JTAC support by opening a case.
You can remove these files manually after taking 1 or 2 files for backup for analysis. There is no problem in remving these files
We cannot configure the jnprsnmpd process to automatically delete old core files, however in SBRC we have an option to discards the oldest core files and generates a new core file if it fails. This is possible by enabling the configuration in sbrd.conf using the parameter ULIMIT_CORE_COUNT. This is applicable only if the radius process fails during the core generation.
Hope this calrifies your query.
Core files are symptomatic of an underlying software defect. You should open a case with JTAC to see whether it's a know issue or not. You can delete older cores if they look like they are the same as the later ones if you need space. But save at least 1 for JTAC.