In article <cam_nd6ovp2_qqvxh70zvhh3aka59kdjc50paorz8mjesjzf...@mail.gmail.com>, Miwa Susumu <miwa...@gmail.com> wrote: >2016-02-07 1:15 GMT+09:00 Christos Zoulas <chris...@astron.com>: >>>>>pflogd process consume CPU. >>>>>Because of that in load average is too high. >>>> Can you ktrace it? >>>this is kdump output. >>>sakura# kdump ktrace.out >>> 974 1 pflogd EMUL "netbsd" >>> 974 1 pflogd CALL read(3,0xbb912000,0x80000) >>> 974 1 pflogd RET read -1 errno 35 Resource temporarily >>> unavailable >> And what does 'fstat -p 974' say about fd 3? > >fd 3 is bpf. > >% fstat -p 974 >USER CMD PID FD MOUNT INUM MODE SZ|DV R/W >_pflogd pflogd 974 root /var 389524 drwxr-xr-x 512 r >_pflogd pflogd 974 wd /var 389524 drwxr-xr-x 512 r >_pflogd pflogd 974 0 / 380813 crw-rw-rw- null rw >_pflogd pflogd 974 1 / 380813 crw-rw-rw- null rw >_pflogd pflogd 974 2 / 380813 crw-rw-rw- null rw >_pflogd pflogd 974 3* bpf rec=0, dr=0, cap=0, pid=1135, >promisc, seesent, idle >_pflogd pflogd 974 4 /var 1855883 -rw------- 24 rw >_pflogd pflogd 974 5* unix stream 0xc4eda9b8 <-> 0xc4edaa08 >
Seems that there is a general problem with non-blocking fd's on 7. Can you start pflogd with ktrace or gdb and see where it is setting non-blocking I/O on the fd? christos
