OPERATIONAL DEFECT DATABASE
...
...
A Nexus switch may see repeated crashes on the NTPd process until the "/var/sysmgr/logs" or "/var/sysmgr/tmp_logs" directory becomes full. This may trigger a reset of the engine or other impact to normal operation. %SYSMGR-2-LAST_CORE_BASIC_TRACE: : PID XXXXX with message ntpd(non-sysmgr) crashed, core will be saved. %SYSMGR-2-LAST_CORE_BASIC_TRACE: : PID XXXXX with message ntpd(non-sysmgr) crashed, core will be saved. %SYSMGR-2-LAST_CORE_BASIC_TRACE: : PID XXXXX with message ntpd(non-sysmgr) crashed, core will be saved. %SYSMGR-3-CORE_OP_FAILED: Core operation failed: copy_cores_from_core_regs_table: Core files: 6 exceeds max: 5 for non-sysmgr Core not saved, will be moved to logflash if exist %SYSMGR-3-VAR_SYSMGR_FULL: System Manager file storage usage is unexpectedly high at 100%. This may impact system functions.
An NTP access-group is configured along with a matching access-list to restrict which devices the switch listens to for NTP updates.
Remove any "ntp access-group" configurations. "ntp commit" may also need to be entered for the new NTP configuration to take effect. Sometimes a reload or supervisor switchover is needed for the workaround to take effect. If NTP packets must be filtered, you can instead do this via a CoPP policy. This will deny the packets before they hit the control plane, meaning they will not be processed by the NTPd process to cause a crash. ================================================== How to Clear Existing Core Files from the "/var/sysmgr/" Directory ================================================== Note that the above will only stop new NTP cores from being written. To clear any existing cores seen via "show system internal dir /var/sysmgr/logs" or via "show system internal dir /var/sysmgr/tmp_logs/", either: 1) Perform a supervisor failover via "redundancy force-switchover" if the switch has dual supervisor engines. 2) Contact the TAC to load a debugging plugin to manually clear these directories. Also note that may also take *multiple attempts* to clear the "/var/sysmgr/logs" and "/var/sysmgr/tmp_logs" directories before the NTP process stops crashing, even after the "ntp access-group" config is removed. You may also see core files for the "non-sysmgr" process as a side effect of this bug. Alternatively, you can clear all core files by "clear cores".
Issue does not exist in releases prior to NX-OS 6.2(16). Therefore, 6.2(14) and earlier are not affected. The next planned release of 6.2 code to have this fix will be 6.2(18) tentatively due on Cisco.com on January or February 2017.
Click on a version to see all relevant bugs
Cisco Integration
Learn more about where this data comes from
Bug Scrub Advisor
Streamline upgrades with automated vendor bug scrubs
BugZero Enterprise
Wish you caught this bug sooner? Get proactive today.
