Martijn van Oosterhout writes:
> It's still a reasonable suggestion. The maximum offset is the number of
> rows in the table. You'll notice when the output is empty.
Once I find the point where the output is empty then what?
> Do you have
> an idea how much data it contains?
Yes. Around 87 million rows.
> If you connect via psql, you can set the client_min_message to DEBUG to
> get some more stuff.
How?
Tried set client_min_message='DEBUG';
> Well, it will stop complaining. What will happen is that any
> transactions involving those transaction IDs will be assumed to have
> been committed.
If the pg_clog files are to keep track of transactions, shouldn't a "pg_ctl
restart" rollback all pending transactions.. so there are no pending
transactions upon the restart and this error should not appear again?
Using 8.1.4
> This may be ok, but in extreme cases it could lead to
> broken constraints.
That exteme case sounds pretty bad. :-(
Will it be safe to do after a restart? After all there should not be any
transactions..
> However, pg_clog/0000 is the very first transaction file created. Is it
> in the range of the files that do do exist?
There are 228 files in the directory and the oldest one is "016E" from about
a month ago.
> Are you sure some other
> process didn't remove it accedently?
Not that I can think off. I didn't even know what this diretory was until
this problem showed up.. much less delete anything from it.