Launching soon: The world's first vendor agnostic bug scrubLearn more & join waitlist
OPERATIONAL DEFECT DATABASE
...
...
This defect documents a rapid memory leak in the "ngiolite" process. This leak can result in the device crashing after approximately 2 hours due to memory exhaustion. The memory leak can be seen using the following commands: # show platform resources **State Acronym: H - Healthy, W - Warning, C - Critical Resource Usage Max Warning Cr --------------------------------------------------------------------------------- RP0 (ok, active) Control Processor 29.17% 100% 80% 90 DRAM 1767MB(95%) 1853MB 88% 93 <<<----- # show processes memory platform sorted System memory: 1898080K total, 1812412K used, 85668K free, Lowest: 79996K Pid Text Data Stack Dynamic RSS Name ---------------------------------------------------------------------- 25137 93 436560 136 44 436560 ngiolite <<<----- 10843 362319 351044 136 156 351044 linux_iosd-imag 14850 12067 83788 136 188 83788 fman_fp_image The device may also exhibit the following memory warning logs: %PLATFORM-4-ELEMENT_WARNING: R0/0: smand: RP/0: Used Memory value % exceeds warning level 88%. Top memory allocators are: Process: ngiolite_cc_0. Tracekey: 1#d78e51be83b7bc5486751adc68d8c456 Callsite ID: 2148339713 (diff_call: 237557).
This issue was found on a C1100TG-1N32A with multiple Async interfaces configured. There are no specific triggers known at this time. This issue appears to happen randomly, and the memory can potentially stabilize.
Reloading the module should recover the used memory: hw-module subslot 0/1 reload force This operation could be automated through EEM applet that will trigger when the memory usage warning is produced, for example: event manager applet IOMD-reload authorization bypass event syslog pattern "%PLATFORM-4-ELEMENT_WARNING" action 0.0 syslog msg "LOW MEMORY DETECTED. Reload the module 0/1 !" action 0.1 cli command "enable" action 1.1 cli command "hw-module subslot 0/1 reload force"
This issue is currently under investigation.