Horton

Re: "unexpected EOF" messages

Tom Lane <tgl@sss.pgh.pa.us>

View thread
From: Tom Lane <tgl@sss.pgh.pa.us>
To: "Kevin Grittner" <Kevin.Grittner@wicourts.gov>
Cc: "Simon Riggs" <simon@2ndquadrant.com>, "Alvaro Herrera" <alvherre@commandprompt.com>, "Robert Haas" <robertmhaas@gmail.com>, "Magnus Hagander" <magnus@hagander.net>, "Pg Hackers" <pgsql-hackers@postgresql.org>
Date: 2012-05-03T18:25:51Z
Lists: pgsql-hackers 
"Kevin Grittner" <Kevin.Grittner@wicourts.gov> writes:
> I still think it might be useful to differentiate in our server log
> between the case where the transaction failed and the case where the
> transaction committed but we don't know that the client got the news
> of that.  How about something like:
 
> 2DP01  connection_lost_during_transaction
> 2DP02  connection_lost_during_commit_notification

That would be a useful distinction, but I'm not sure how easily our
code can make it.
 
> I'm less sure what makes sense if the connection fails while idle
> (not in transaction).  If you don't like "Class 08 * Connection
> Exception" for that, I'm not quite sure where it belongs.

I'm not convinced that these cases belong in any of the standard's
classes.  IMO the standard is only standardizing application-visible
error cases, which these are not.  In particular I think class 2D is
not appropriate, since AFAICS the standard means that to pertain to
incorrect issuance of a COMMIT or ROLLBACK command.

			regards, tom lane