Graeme started a short thread on databases in "The Other Place" a few
days ago, but I thought this might be of sufficient general relevance to
raise here.
I had a system outage this morning, with all apps suddenly losing their
connectivity to the PostgreSQL server.
It turned out that the cause of that was that Debian had done an
unattended upgrade of the Postgres server, and by restarting it had
killed all persistent connections. There is no "Can we kill your
database when we feel like it?" question during Debian installation.
I anticipate that the same problem will affect other databases or
software to which a client program maintains a persistent session,
unless explicit steps are taken to recognise and recover from a
server-side restart.
Noting that the traditional way of using the data-aware controls
introduced by Delphi etc., is particularly vulnerable, and noting that
the FPC/Lazarus controls do a good job of presenting a common API
irrespective of what backend server is being used, would it be feasible
to have a "reconnect monitor" or similar to help recover from this sort
of thing?
--
Mark Morgan Lloyd
markMLl .AT. telemetry.co .DOT. uk
[Opinions above are the author's, not those of his employers or colleagues]
_______________________________________________
fpc-pascal maillist - fpc-pascal@lists.freepascal.org
http://lists.freepascal.org/cgi-bin/mailman/listinfo/fpc-pascal
