Bug#633793: libc6: [sparc] getaddrinfo() times out, succeeds on retry
Package: libc6
Version: 2.11.2-10
Severity: important
getaddrinfo() randomly encounters 5 to 10 second delays in DNS name
resolution, and sometimes fails. Using strace, I was able to track this
down to timeouts in waiting for the UDP socket to be ready for a read().
Adjusting the timeout or attempts options in /etc/resolv.conf influences
the timeout behavior.
Running tcpdump on the client shows that the DNS server is sending
responses to the client's queries, but the resolver seems to ignore them.
When running strace on the following python script:
import socket
socket.getaddrinfo('www.google.com', 80)
I get the following output from strace (note the Timeouts):
0.000298 socket(PF_INET, SOCK_DGRAM|SOCK_NONBLOCK, IPPROTO_IP) = 3
0.000213 connect(3, {sa_family=AF_INET, sin_port=htons(53), sin_addr=inet_addr("130.127.255.250")}, 16) = 0
0.000251 gettimeofday({1310583673, 595304}, NULL) = 0
0.000180 poll([{fd=3, events=POLLOUT}], 1, 0) = 1 ([{fd=3, revents=POLLOUT}])
0.000231 send(3, "D\230\1\0\0\1\0\0\0\0\0\0\3www\6google\3com\0\0\1\0\1", 32, MSG_NOSIGNAL) = 32
0.000307 poll([{fd=3, events=POLLIN|POLLOUT}], 1, 5000) = 1 ([{fd=3, revents=POLLOUT}])
0.000239 send(3, "\237K\1\0\0\1\0\0\0\0\0\0\3www\6google\3com\0\0\34\0\1", 32, MSG_NOSIGNAL) = 32
0.000303 gettimeofday({1310583673, 596567}, NULL) = 0
0.000262 poll([{fd=3, events=POLLIN}], 1, 4998) = 1 ([{fd=3, revents=POLLIN}])
0.000453 ioctl(3, 0x4004667f, 0xff98bc50) = 0
0.000287 recvfrom(3, "D\230\201\200\0\1\0\7\0\4\0\4\3www\6google\3com\0\0\1\0\1"..., 2048, 0, {sa_family=AF_INET, sin_port=htons(53), sin_addr=inet_addr("130.127.255.250")}, [16]) = 284
0.000538 gettimeofday({1310583673, 598108}, NULL) = 0
0.000174 poll([{fd=3, events=POLLIN}], 1, 4997) = 0 (Timeout)
5.002406 gettimeofday({1310583678, 600685}, NULL) = 0
0.000173 poll([{fd=3, events=POLLOUT}], 1, 0) = 1 ([{fd=3, revents=POLLOUT}])
0.000225 send(3, "D\230\1\0\0\1\0\0\0\0\0\0\3www\6google\3com\0\0\1\0\1", 32, MSG_NOSIGNAL) = 32
0.000301 poll([{fd=3, events=POLLIN}], 1, 5000) = 1 ([{fd=3, revents=POLLIN}])
0.001226 ioctl(3, 0x4004667f, 0xff98bc50) = 0
0.000359 recvfrom(3, "D\230\201\200\0\1\0\7\0\4\0\4\3www\6google\3com\0\0\1\0\1"..., 2048, 0, {sa_family=AF_INET, sin_port=htons(53), sin_addr=inet_addr("130.127.255.250")}, [16]) = 284
0.000531 gettimeofday({1310583678, 603502}, NULL) = 0
0.000210 poll([{fd=3, events=POLLOUT}], 1, 4997) = 1 ([{fd=3, revents=POLLOUT}])
0.000239 send(3, "\237K\1\0\0\1\0\0\0\0\0\0\3www\6google\3com\0\0\34\0\1", 32, MSG_NOSIGNAL) = 32
0.000296 gettimeofday({1310583678, 604250}, NULL) = 0
0.000176 poll([{fd=3, events=POLLIN}], 1, 4996) = 0 (Timeout)
5.001422 close(3) = 0
Hard-coding an entry for www.google.com into /etc/hosts results in no
timeout from getaddrinfo().
Additionally, gethostbyname() does not seem to result in any delays
regardless of whether or not the host is in /etc/hosts or is coming from
DNS.
This problem is not specific to python. We first noticed the problem in
the LDAP client, but were able to track it down to a general problem with
getaddrinfo().
-- System Information:
Debian Release: 6.0.2
APT prefers stable-updates
APT policy: (500, 'stable-updates'), (500, 'stable')
Architecture: sparc (sparc64)
Kernel: Linux 2.6.32 (SMP w/1 CPU core)
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Versions of packages libc6 depends on:
ii libc-bin 2.11.2-10 Embedded GNU C Library: Binaries
ii libgcc1 1:4.4.5-8 GCC support library
libc6 recommends no packages.
Versions of packages libc6 suggests:
ii debconf [debconf-2.0] 1.5.36.1 Debian configuration management sy
pn glibc-doc <none> (no description available)
ii locales 2.11.2-10 Embedded GNU C Library: National L
-- debconf information:
glibc/upgrade: true
glibc/restart-services:
glibc/disable-screensaver:
glibc/restart-failed:
Reply to: