There's some sort of race condition with the new FSM stuff

Two different buildfarm machines are currently showing the same failure:
http://www.pgbuildfarm.org/cgi-bin/show_log.pl?nm=kudu&dt=2008-10-13%2015:30:00
http://www.pgbuildfarm.org/cgi-bin/show_log.pl?nm=dragonfly&dt=2008-10-13%2016:30:01

The postmaster log in each case shows

ERROR: could not fsync segment 0 of relation 1663/16384/29270/1: No such file or directory
ERROR: checkpoint request failed
HINT: Consult recent messages in the server log for details.
STATEMENT: DROP TABLESPACE testspace;
ERROR: could not fsync segment 0 of relation 1663/16384/29270/1: No such file or directory
ERROR: checkpoint request failed
HINT: Consult recent messages in the server log for details.
STATEMENT: DROP TABLESPACE testspace;

which looks like an fsync request has been allowed to hang around too
long, ie, after its file is already gone. Since /1 indicates the FSM
fork, I suppose this has been introduced by the new FSM code. I haven't
dug any more deeply than that though.

regards, tom lane