The problem:
1) I create a stupid function like:
CREATE FUNCTION p4() RETURNS integer AS '
BEGIN
FOR i IN 1..1000000 LOOP
insert into info values (random(),random());
END LOOP;
RETURN 1;
END;
' LANGUAGE plpgsql;

2) I create a new db and run "select p4();"

3) I open three connection to the same db and start operation as
"begin work; select p4(); update info set data=RANDOM();commit;"
"create index gino on info(x);"
"begin work; update info set data=RANDOM();"

Well, db schema is nothink more than this:
create table info(meta text, data text, id serial primary key);

4) after a lot of work/insert I've got this:
LOG: recycled transaction log file "00000002000000B0"
LOG: recycled transaction log file "00000002000000B2"
LOG: recycled transaction log file "00000002000000B3"
ERROR: could not access status of transaction 859388209
DETAIL: open of file "/home/gelma/pg4/bin/../db/pg_clog/0333" failed:
No such file or directory
ERROR: could not access status of transaction 859388209
DETAIL: open of file "/home/gelma/pg4/bin/../db/pg_clog/0333" failed:
No such file or directory
siae=# ERROR: could not access status of transaction 859388209
DETAIL: open of file "/home/gelma/pg4/bin/../db/pg_clog/0333" failed:
No such file or directory

I can reproduce this over and over...

The real problem is my configuration: a linux unstable kernel with a debian
unstable branch in a shiny new laptop (with no more test than a memtest86).
Everything else seems working well, of course it doesn't mean it's not my
fault.
Anyway, I write to you asking:
a) do you think it could be a problem of 7.4beta2 (not a problem of mine)?
I remember Tom solved something like this years ago.
b) if so, can you show me next steps to narrow down the problem to give you
a little help?

Thanks a lot for your work,
Andrea Gelmini

---------------------------(end of broadcast)---------------------------
TIP 2: you can get off all lists at once with the unregister command
(send "unregister YourEmailAddressHere" to majordomo (AT) postgresql (DOT) org)

andrea gelmini <andrea.gelmini (AT) linux (DOT) it> writes:
Uh, there's no "x" column in "info".

I spent some time trying to reproduce this, with no success. Could you
give a clearer recipe?

regards, tom lane

---------------------------(end of broadcast)---------------------------
TIP 3: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to majordomo (AT) postgresql (DOT) org so that your
message can get through to the mailing list cleanly