(I've rebooted after installing the "7.1-RELEASE-p5" patch with freebsd-update. It doesn't seem that this patch don't change the default kernel, that's why uname says 7.1-RELEASE-p4 and not 7.1-RELEASE-p5) The commands host, dig and nslookup hang into the _umtx_op system call (see attached ktrace for host). They can only be killed with SIGKILL. However, any others tools can resolve address using DNS (ping, tcpdump). Rollbacking to 7.1-RELEASE-p4 (freebsd-update rollback) suppressed the problem. Searching the web I saw (may be linked to my issue) : - bug report bin/85810 "dig(1): nslookup/dig hang and processes cannot be killed" - http://www.mailinglistarchive.com/freebsd-questions@freebsd.org/msg50315.html As said by freebsd-update, 7.1-RELEASE-p5 changes : /lib/libc.so.7 /lib/libcrypto.so.5 /rescue/* /sbin/init /usr/bin/dig /usr/bin/host /usr/bin/nslookup /usr/bin/nsupdate /usr/bin/openssl /usr/lib/libc.a /usr/lib/libc_pic.a /usr/lib/libcrypto.a /usr/lib/libssl.a /usr/lib/libssl.so.5 /usr/libexec/lukemftpd /usr/libexec/telnetd /usr/sbin/dnssec-keygen /usr/sbin/dnssec-signzone /usr/sbin/lwresd /usr/sbin/named /usr/sbin/named-checkconf /usr/sbin/named-checkzone /usr/sbin/named-compilezone /usr/sbin/ntpd /usr/sbin/rndc-confgen Fix: Patch attached with submission follows: How-To-Repeat: run : host freebsd.org
I've run into this as well, but only within jail environments. The host environment does not deadlock, but making these calls inside jails *does* deadlock on 7.1-RELEASE-p5. Rolling back to 7.1-RELEASE fixed it for me.
I've had the same problem not in a jail. I first noticed it running portsnap fetch and freebsd-update fetch. I did a roll back and the problem was gone.
Is this still a problem on later releases? I'm guessing this ticket can be closed.
Bug does not exist anymore in recent versions.