On 07.03.2012 12:53, Simon Slavin wrote:
On 7 Mar 2012, at 11:22am, Marcus Grimm<mgr...@medcom-online.de> wrote:
I evaluating the need to issue a (timeconsuming) "pragma integrity_check" prior
running the backup to avoid that a backup will be overwritten
with an invalid database.
Did you know about
<http://www.sqlite.org/pragma.html#pragma_quick_check>
? It misses out a check, but the check it misses out is for data which can be
recovered simply by recreating corrupt indexes.
Yes, I'm aware about that check.
On the other hand, if you are using this as a warning system that corruption
has occurred and therefore you may have faulty hardware, then the more thorough
the check, the better.
Yes, it is just about safe design. I'm currently reviewing my implementation
and thought
about the unlikely event of a DB corruption - In that case my current backup
mechanism will become easily useless, depending if the corruption is so
small that it is not detected during the normal program flow.
(The application is a server that runs for weeks)
However, it is so far a theoretical issue - we haven't yet seen any corruption
in the field. :-)
Thanks also to Dan for the clarification.
Kind regards
Marcus
Simon.
_______________________________________________
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
_______________________________________________
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
