Leonard Meyer <lmeyer@excilys.com> wrote:
> 2014-12-04 16:30 GMT+01:00 Kevin Grittner <kgrittn@ymail.com>:
>> Leonard Meyer <lmeyer@excilys.com> wrote:
>>> We're doing an ETL Akka application using PostgreSQL 9.3.5
>>> using driver 9.3-1102-jdbc41. We're currently testing our
>>> application resilience by shutting down the database in the
>>> middle of our processing.
>>
>> *Exactly* how are you "shutting down the database"?
> We call stop on the postgresql service so "sudo service
> postgresql stop".
Hmm. That shouldn't prevent the client side from seeing the
connection as closed (as would happen if the network was unplugged
or the server abruptly turned off) -- so the TCP keepalive settings
should not come into play. Given that, it does sound like a likely
bug in the JDBC driver. The best place to report that is the
pgsql-jdbc list; I don't think that the people who do most of the
work on the JDBC driver tend to read this pgsql-bug list. Please
post there, and include that stack trace again.
--
Kevin Grittner
EDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company