[nycbug-talk] postgres incremental backups?

[Francisco Reyes]

Peter Wright writes:


> getting his code to play nice with 8.x.  i'm sure he was vacuuming/etc as
> we talked about that

Did he ever check the vacuum worked. :-)
It is possible to vacuum and to have it fail. If you go over lots of 
transactions and you have default seetings.. there is a resourse (FSM pages) 
which needs to be increased.

> so for lack of further evidence i'm guessing his
> code was doing something nasty (at the time of the bug report it was
> causing a deadlock for example).

What language was the program?
I have been using postgresq for something like 5+ years.. and never had any 
problems until I tried a program called dspam (writen in c). That program 
caused me more database problems (including corrupting) in a few weeks than 
I had with Postgresql in all the time I have been using it. When I wrote to 
the dspam list they basically replied it was postgresql's fault.. Possible, 
but I was not about to keep trashing my DB to figure out why this one 
program caused so many problems.