https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=279742
--- Comment #3 from Konstantin Belousov <[email protected]> --- (In reply to dgilbert from comment #2) Look at the vmstat man page, the -M flag. The vmstat -z was included into the core.txt, but vmstat -m failed. In vmstat -z, there is outstanding number of vm objects (184079069), which more or less aligns with the number of allocated tmpfs nodes. This, together with 64/32/16 malloc zones, seems to eat all machine memory: (gdb) p (264 + 232 + 64 + 32 + 16)*184168869ull $3 = 111974672352 I wonder if it is 'simply' tmpfs metadata consuming memory. Try to limit tmpfs appetite perhaps? -- You are receiving this mail because: You are the assignee for the bug.
