[Kea-users] DHCP-DDNS server repeatedly logging "Resource temporarily unavailable"

[Derek Lambert]

Hi,

On Fedora 23 running Kea 1.0.0-1.fc23 x86_64 installed from Fedora repo.

When the DHCP-DDNS server is started it repeatedly logs "Resource
temporarily unavailable" errors:

2016-02-04 07:04:44.734 INFO  [kea-dhcp-ddns.dhcpddns/11234]
DHCP_DDNS_STARTED Kea DHCP-DDNS server version 1.0.0 started
2016-02-04 07:04:44.734 ERROR [kea-dhcp-ddns.libdhcp-ddns/11234]
DHCP_DDNS_NCR_UDP_RECV_ERROR UDP socket receive error while listening for
DNS Update requests: Resource temporarily unavailable
2016-02-04 07:04:44.734 ERROR [kea-dhcp-ddns.dhcp-to-d2/11234]
DHCP_DDNS_QUEUE_MGR_RECV_ERROR application's queue manager was notified of
a request receive error by its listener.
2016-02-04 07:04:44.734 INFO  [kea-dhcp-ddns.dhcpddns/11234]
DHCP_DDNS_QUEUE_MGR_RECOVERING application is attempting to recover from a
queue manager IO error
2016-02-04 07:04:44.735 ERROR [kea-dhcp-ddns.libdhcp-ddns/11234]
DHCP_DDNS_NCR_UDP_RECV_ERROR UDP socket receive error while listening for
DNS Update requests: Resource temporarily unavailable
2016-02-04 07:04:44.735 ERROR [kea-dhcp-ddns.dhcp-to-d2/11234]
DHCP_DDNS_QUEUE_MGR_RECV_ERROR application's queue manager was notified of
a request receive error by its listener.
2016-02-04 07:04:44.735 INFO  [kea-dhcp-ddns.dhcpddns/11234]
DHCP_DDNS_QUEUE_MGR_RECOVERING application is attempting to recover from a
queue manager IO error
2016-02-04 07:04:44.735 ERROR [kea-dhcp-ddns.libdhcp-ddns/11234]
DHCP_DDNS_NCR_UDP_RECV_ERROR UDP socket receive error while listening for
DNS Update requests: Resource temporarily unavailable
2016-02-04 07:04:44.735 ERROR [kea-dhcp-ddns.dhcp-to-d2/11234]
DHCP_DDNS_QUEUE_MGR_RECV_ERROR application's queue manager was notified of
a request receive error by its listener.
2016-02-04 07:04:44.735 INFO  [kea-dhcp-ddns.dhcpddns/11234]
DHCP_DDNS_QUEUE_MGR_RECOVERING application is attempting to recover from a
queue manager IO error

The log file fills quickly, and within a minute is hundreds of megabytes.

Changing the DDNS log level in the config to fatal suppresses these
messages, but the process shows 100% CPU usage. SELinux is configured in
permissive mode, and there is no other process listening on port 53001. The
DDNS server also exhibits this behavior on a fresh Fedora 23 install with
no modifications to the system or configuration.

An strace of the process shows the following calls:

0.000067 socket(PF_INET, SOCK_DGRAM, IPPROTO_UDP) = 10<UDP:[543400]>
0.208351 epoll_ctl(8<anon_inode:[eventpoll]>, EPOLL_CTL_ADD,
10<UDP:[543400]>, {EPOLLIN|EPOLLPRI|EPOLLERR|EPOLLHUP|EPOLLET,
{u32=788755824, u64=94258141033840}}) = 0
0.000299 setsockopt(10<UDP:[543400]>, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
0.000111 bind(10<UDP:[543400]>, {sa_family=AF_INET, sin_port=htons(53001),
sin_addr=inet_addr("127.0.0.1")}, 16) = 0
0.000351 ioctl(10<UDP:[127.0.0.1:53001]>, FIONBIO, [1]) = 0
0.000165 recvmsg(10<UDP:[127.0.0.1:53001]>, 0x7ffd3c3c0c40, 0) = -1 EAGAIN
(Resource temporarily unavailable)
I've also opened Fedora bug
https://bugzilla.redhat.com/show_bug.cgi?id=1301673 for this issue. Asking
here in case I'm overlooking a simple solution or this is a bug in Kea.

Thanks!

Derek Lambert
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.isc.org/pipermail/kea-users/attachments/20160204/5dc21a95/attachment.htm>