Hi,
After upgrading a NetBSD-5 system to 6 RC2, I find that named is using 100%
CPU. The symptoms look identical to this report:
http://mail-index.netbsd.org/current-users/2012/06/17/msg020440.html
My ktrace shows a continuous stream of:
11333 6 named CALL __gettimeofday50(0xbabff87c,0)
11333 6 named RET __gettimeofday50 0
11333 6 named CALL ___lwp_park50(0xbabff87c,0,0xbb51b020,0xbb51b020)
11333 4 named CALL __gettimeofday50(0xbafffa1c,0)
11333 4 named RET __gettimeofday50 0
11333 4 named CALL __gettimeofday50(0xbafff91c,0)
11333 4 named RET __gettimeofday50 0
11333 4 named CALL _lwp_unpark(6,0xbb51b020)
11333 4 named RET _lwp_unpark 0
11333 6 named RET ___lwp_park50 0
I do not see any resolution to the problem in the archives but I find that
commenting out the
dnssec-lookaside auto;
option in the named.conf stops the high CPU usage. Is this a reasonable
workaround? Is there a better way to resolve this?
Regards,
Sverre
PS I see the same behavior with NetBSD-current (from late August).