Query timeouts

[Kyle Gordon]

Hi all,

I'm having a few problems with Bind9 (9.2.3+9.2.4-rc5-1) on Debian =
testing (although the same thing happens with the version from Debian =
stable)

It seems to be timing out either shortly after starting up, or about 6 =
days of normal running. I have tried running it through strace, and I =
get the following output.=20

[pid 25109] gettimeofday( <unfinished ...>
[pid 31874] futex(0x8096530, FUTEX_WAKE, 1 <unfinished ...>
[pid 25109] <... gettimeofday resumed> {1090793595, 142561}, NULL) =3D 0
[pid 31874] <... futex resumed> )       =3D 0
[pid 25109] futex(0x8096560, FUTEX_WAIT, 2, NULL <unfinished ...>
[pid 31874] futex(0x8096560, FUTEX_WAKE, 1 <unfinished ...>
[pid 25109] <... futex resumed> )       =3D -1 EAGAIN (Resource =
temporarily unavailable)
[pid 31874] <... futex resumed> )       =3D 0
[pid 25109] futex(0x8096570, FUTEX_WAKE, 1 <unfinished ...>
[pid 31874] clock_gettime(0,  <unfinished ...>
[pid 25109] <... futex resumed> )       =3D 0
[pid 31874] <... clock_gettime resumed> {1090793595, 145303000}) =3D 0
[pid 25109] futex(0x8096560, FUTEX_WAKE, 1 <unfinished ...>
[pid 31874] futex(0x8096570, FUTEX_WAIT, 27, {2571, 329731000} =
<unfinished ...>
[pid 25109] <... futex resumed> )       =3D 0
[pid 31874] <... futex resumed> )       =3D -1 EAGAIN (Resource =
temporarily unavailable)
[pid 25109] gettimeofday( <unfinished ...>
[pid 31874] gettimeofday( <unfinished ...>
[pid 25109] <... gettimeofday resumed> {1090793595, 147251}, NULL) =3D 0
[pid 31874] <... gettimeofday resumed> {1090793595, 147895}, NULL) =3D 0
[pid 25109] gettimeofday( <unfinished ...>
[pid 31874] futex(0x8096530, FUTEX_WAKE, 1 <unfinished ...>
[pid 25109] <... gettimeofday resumed> {1090793595, 148586}, NULL) =3D 0
[pid 31874] <... futex resumed> )       =3D 0
[pid 25109] futex(0x8096560, FUTEX_WAIT, 2, NULL <unfinished ...>
[pid 31874] futex(0x8096560, FUTEX_WAKE, 1 <unfinished ...>
[pid 25109] <... futex resumed> )       =3D -1 EAGAIN (Resource =
temporarily unavailable)
[pid 31874] <... futex resumed> )       =3D 0
[pid 25109] futex(0x8096570, FUTEX_WAKE, 1 <unfinished ...>
[pid 31874] clock_gettime(0,  <unfinished ...>
[pid 25109] <... futex resumed> )       =3D 0
[pid 31874] <... clock_gettime resumed> {1090793595, 151301000}) =3D 0
[pid 25109] futex(0x8096560, FUTEX_WAKE, 1 <unfinished ...>
[pid 31874] futex(0x8096570, FUTEX_WAIT, 28, {2571, 323733000} =
<unfinished ...>
[pid 25109] <... futex resumed> )       =3D 0
[pid 31874] <... futex resumed> )       =3D -1 EAGAIN (Resource =
temporarily unavailable)
[pid 25109] time( <unfinished ...>
[pid 31874] gettimeofday( <unfinished ...>
[pid 25109] <... time resumed> [1090793595]) =3D 1090793595
[pid 31874] <... gettimeofday resumed> {1090793595, 154628}, NULL) =3D 0
[pid 25109] getpid( <unfinished ...>
[pid 31874] futex(0x8096530, FUTEX_WAKE, 1 <unfinished ...>
[pid 25109] <... getpid resumed> )      =3D 27793
[pid 31874] <... futex resumed> )       =3D 0
[pid 25109] rt_sigaction(SIGPIPE, {0x4f0b2470, [], 0},  <unfinished ...>
[pid 31874] clock_gettime(0,  <unfinished ...>
[pid 25109] <... rt_sigaction resumed> {SIG_IGN}, 8) =3D 0
[pid 31874] <... clock_gettime resumed> {1090793595, 158816000}) =3D 0
[pid 25109] send(3, "<30>Jul 25 23:13:15 named[27793]"..., 41, 0 =
<unfinished ...>
[pid 31874] futex(0x8096570, FUTEX_WAIT, 29, {2571, 316218000}

I've gone as far as binning all my bind9 config files (but not my =
zonefiles), and starting them from scratch again but the problem =
persists. The system date appears to be correct, and is synchronised =
with ntpdate and ntp.

I've had look with Google for any ideas relating to gettimeofday and =
clock_gettime in the context of named, but I don't see anything that =
helps.

Anyone have any thoughts?

Many thanks

Kyle