Janning Vygen <vygen@kicktipp.de> writes:
> I am investigating some error messages in my log file:
> Apr 15 08:04:34 postgres[20686]: [4-1] 2009-04-15 08:04:34 CEST ERROR:
> invalid byte sequence for encoding "UTF8": 0x81
> Apr 15 08:04:34 postgres[20686]: [4-2] 2009-04-15 08:04:34 CEST HINT: This
> error can also happen if the byte sequence does not match the encoding
> expected by the
> Apr 15 08:04:34 postgres[20686]: [4-3] server, which is controlled by
> "client_encoding".
> I want to know which statement produces this error. I have
> show log_min_error_statement ;
> error
> show log_error_verbosity;
> default
> so i think the error statement should be logged. Why is the error statement
> not logged?
Because the error is happening while trying to convert the statement
into the database's encoding. We cannot log the string we have without
creating an encoding mismatch in the postmaster log.
> I always see 0x81 and 0xe46973 as invalid byte sequence. Can someone give me a
> hint what characters in which encoding those bytes might be?
LATIN1, or some other one of the single-byte LATINn encodings, likely.
regards, tom lane
