Re: Suggestion for To Do List - Client timeout please.

Tom Lane writes:

> There is something wrong with your system, not with Postgres. Any
> reasonable TCP stack will time out within circa 1 minute if no response.
>
> Example (sss is a machine on my LAN that's not presently up):
>
> $ time psql -h sss
> psql: PQconnectPoll() -- connect() failed: Connection timed out
> Is the postmaster running (with -i) at 'sss'
> and accepting connections on TCP/IP port '5432'?
>
> real 1m14.27s
> user 0m0.01s
> sys 0m0.01s
> $
>
> This particular timeout length is probably specific to HPUX, but the
> point is that you have a local system problem.

I can observe something peculiar:

With current sources:

peter ~$ time pg-install/bin/psql -h ralph
psql: could not connect to server: No route to host
Is the server running on host ralph and accepting
TCP/IP connections on port 5432?

real 0m3.013s
user 0m0.010s
sys 0m0.010s

With 7.0.2:

peter ~$ time psql -h ralph
psql: PQconnectPoll() -- connect() failed: No route to host
Is the postmaster running (with -i) at 'ralph'
and accepting connections on TCP/IP port '5432'?

[hangs]

The backtrace shows:

#0 0x401d9a0e in __select () from /lib/libc.so.6
#1 0x4002f3b0 in b2c3 () from /usr/lib/libpq.so.2.1
#2 0x4002666e in pqFlush () from /usr/lib/libpq.so.2.1
#3 0x40022bc2 in closePGconn () from /usr/lib/libpq.so.2.1
#4 0x40022c67 in PQfinish () from /usr/lib/libpq.so.2.1
#5 0x805167d in main ()

I suspect that this may be because of the questionable TCP implementation
in Linux that you argued about with Alan Cox et al. a while ago, though I
don't pretend to fathom the details. Apparently something in libpq
changed in between, however.

But that reinforces your point that "something is wrong with your system".
;-)

--
Peter Eisentraut peter_e(at)gmx(dot)net http://funkturm.homeip.net/~peter