Some time ago I reported bug 6291[0], which reported a Xid wraparound,
both as reported in pg_controldata and by txid_current_snapshot.
Unfortunately, nobody could reproduce it.
Today, the same system of ours just passed the wraparound mark
successfully at this time, incrementing the epoch. However, two
standbys have not done the same: they have wrapped to a low txid. At
this time, pg_controldata does report the correct epoch, as I read it,
unlike the original case.
I have not yet tried to reproduce this in a minimal way, but I wanted
to relate this information as soon as possible.
These systems are 9.0.6, on Ubuntu 10.04 LTS, amd64.
[0]: http://archives.postgresql.org/pgsql-bugs/2011-11/msg00094.php
--
fdr