On 11/16/2010 05:15 PM, Tom Lane wrote:
> Keep in mind that these tables are *not* going to survive any type of
> backend crash. Maybe my perceptions are colored because I deal with
> Postgres bugs all the time, but I think of backend crashes as pretty
> common, certainly much more common than an OS-level crash. I'm afraid
> you may be expecting unlogged tables to be significantly more robust
> than they really will be.
But an individual backend crash != server restart, unless that's changed
since 8.1 (yes, I'm still stuck on 8.1 :( )... So if I, for example,
kill -9 a backend that's busy updating a nonlogged table, the table
could be corrupted, but it wouldn't be truncated (and could cause
trouble) for possibly weeks until the postmaster is restarted.
Conversely, even if no backend crash occurs whatsoever, all the
nonlogged tables would be truncated after an orderly postmaster restart.
Just doesn't make sense to me.
-Glen