Aug 2012 16:32:51 -0500
Everyone,
I have two x86_64 servers with CentOS 6.3 that seem to have some
difficulty with the last version of bind :
32:bind-9.8.2-0.10.rc1.el6_3.2.x86_64
When I run htop I am getting a CPU usage of 70 to 80% all of the time.
I am not seeing problems identified in the log files on either machine.
This may have been present before the update of 9.8.2.0, but I certainly
did not notice it. Has anyone else had this problem.
Any solutions??
Greg Ennis
-----------------------------------------------------------------------------
I have continued to have problems with bind's consumption of CPU time.
When I run "strace -p PID#" on the affected PID I get the following that
is generated over and over. This is occurring on both machines. I am
not sure what this means at this point. If any of you can offer
assistance it would be greatly appreciated.
Thanks,
Greg
futex(0x7ff2ae5e808c, FUTEX_WAKE_OP_PRIVATE, 1, 1, 0x7ff2ae5e8088, {FUTEX_OP_SET, 0, FUTEX_OP_CMP_GT, 1}) = 1
futex(0x7ff2ae5e8028, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0x7ff2ae5e9028, FUTEX_WAKE_PRIVATE, 1) = 0
futex(0x7ff2ae5e9078, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0x7ff2ae5e907c, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, 1738218731, {1344789447, 283018000}, ffffffff) = -1 EAGAIN (Resource temporarily unavailable)
gettimeofday({1344788651, 464093}, NULL) = 0
futex(0x7ff2ae5e808c, FUTEX_WAKE_OP_PRIVATE, 1, 1, 0x7ff2ae5e8088, {FUTEX_OP_SET, 0, FUTEX_OP_CMP_GT, 1}) = 1
----------------------------------------------------------------------------------------------------
Here is the fix for anyone else having the same problem :
Take a look at this link for some background info :
https://bugzilla.redhat.com/show_bug.cgi?id=709205
My problem was fixed by doing the following :
chown -R named.named /var/named/chroot
restorecon -R /var/named/chroot/
Greg