4,245 reputation
421
bio website manitou-mail.org/about-dv.php
location Paris, France
age
visits member for 2 years, 8 months
seen 9 hours ago

Let's try to share the bits of knowledge gained during those innumerable hours of programming... Nowadays I use mostly PostgreSQL, C++/Qt, Perl, php on Unix servers.


2d
comment postgresql - a local replication entry doesn't work
Unix domain sockets are refered to with local and never localhost. Anyway the question never mentions localhost in the first place, so it's not obvious why it would even be involved here.
Oct
28
reviewed Approve suggested edit on SQL server transaction log recovery
Oct
26
reviewed Approve suggested edit on TYPE_SYM syntax error when running the mysql command to create or replace type
Oct
24
comment Postgresql single base on multiple HDD partitions
OK. I'd symlink large directories and tranfer their contents to another partition, with pg_xlog being the first candidate if it's big. Once the db is running again, analyze how it's growing and figure a plan to scale using tablespaces.
Oct
24
comment Postgresql single base on multiple HDD partitions
Could you run this at the top of the data directory: du -sh base/* pg_xlog/ and add the results to the question?
Oct
24
answered How to import large sql file in postgres?
Oct
22
answered Why Postgresql 9.2 word boundary not working!
Oct
18
answered Postgres blocking UPDATE on new column?
Oct
13
reviewed Approve suggested edit on Can't execute a MySQL stored procedure from Java
Oct
5
reviewed Approve suggested edit on mssql restore dabase is terminating abnormally
Sep
26
comment How much should I care about prepared statements for simple queries?
It's dubious that the first statement is prepared, because the plan should show: Filter: (account_id = $1). The point of a prepared statement is that the parameters values are not known at prepare/planning time.
Sep
26
answered Format timezone to +-#### instead of +-##
Sep
24
awarded  Autobiographer
Sep
22
answered PostgreSQL 9.3: Index on to_char()
Sep
15
comment Temporary schema per connection?
@Gili: Note that this technique of never committing CREATE SCHEMA is the only one that can guarantee to have them disappear when the unit test gets killed.
Sep
6
answered SQL order by query results in any arbitrary (but reproducible) manner
Aug
30
comment How do I maintain high INSERT-performance on PostgreSQL
"full vacuum" is ambiguous. It may mean VACUUM FULL on a specific table or simple-VACUUM on the whole database or VACUUM FULL on the whole database. Anyway the fact that it helps with performance is suspicious. VACUUM reclaims dead rows resulting from UPDATEs and DELETEs, it's not needed in an INSERT-only scenario.
Aug
28
comment Postgresql 9.1 add nullable column costs more than 10 minutes
Try the query at wiki.postgresql.org/wiki/Lock_dependency_information for a better view. Either you have lingering transactions that forget to commit, or heavy activity with this table that keeps it always busy.
Aug
28
comment Postgresql 9.1 add nullable column costs more than 10 minutes
Before suspecting that it's performing slowly, you need to make sure that the ALTER TABLE is not just waiting for a lock. Mention it in the question if you have checked.
Aug
27
answered postgresql cannot drop a table when autovacuum holds lock on it