Postgresql – External size of table still not released (as disk space) after issuing truncate command

disk-spacepostgresqlpostgresql-9.3truncate

In PostgreSQL 9.3.10 (Ubuntu 14.04 OS), I have a table called groupseen whose size is as follows:

              Table               |  Size   | External Size 
----------------------------------+---------+---------------
          links_groupseen         | 394 MB  | 362 MB

I understand that External Size is the size that related objects of this table (like indices) take.

Upon issuing the command truncate links_groupseen;, only 394 MB disk space was released (I checked right after, via df -h). My two questions are:

What can I do to get rid of the space taken up by this table's indices (External Size) as well? Shouldn't truncate have taken care of everything?

I was using the following command to check the size of this table. Even though External Size hasn't yet been released, using this command now doesn't show the outstanding 362 MB. How do I monitor the outstanding External Size related to this table?

SELECT
   relname as "Table",
   pg_size_pretty(pg_total_relation_size(relid)) As "Size",
   pg_size_pretty(pg_total_relation_size(relid) - pg_relation_size(relid)) as "External Size"
 FROM pg_catalog.pg_statio_user_tables
 ORDER BY pg_total_relation_size(relid) DESC;

In case required, here's how the table looks like:

                            Table "public.links_groupseen"
     Column     |           Type           |                          Modifiers                           
----------------+--------------------------+--------------------------------------------------------------
 id             | integer                  | not null default nextval('links_groupseen_id_seq'::regclass)
 seen_user_id   | integer                  | not null
 seen_at        | timestamp with time zone | not null
 which_reply_id | integer                  | not null
Indexes:
    "links_groupseen_pkey" PRIMARY KEY, btree (id)
    "links_groupseen_seen_user_id" btree (seen_user_id)
    "links_groupseen_which_reply_id" btree (which_reply_id)
Foreign-key constraints:
    "links_groupseen_seen_user_id_fkey" FOREIGN KEY (seen_user_id) REFERENCES auth_user(id) DEFERRABLE INITIALLY DEFERRED
    "links_groupseen_which_reply_id_fkey" FOREIGN KEY (which_reply_id) REFERENCES links_reply(id) DEFERRABLE INITIALLY DEFERRED

Note: this SO answer doesn't solve my problem because the OP in the question stated how disk space wasn't released for him at all. In my case, the subtlety is that only External Size is outstanding.

Best Answer

I think with a truncate even without a VACUUM FULL the space is released back to the OS. My guess is per the other question that you've got a backend process holding the underlying file open.

Find out what the file is. SELECT pg_relation_filepath('links_groupseen');
Find the data directory SHOW data_directory;
Run truncate.
See what's holding it open. lsof $data_directory/$relation_filepath

I think the most likely bet is something is holding it open if not, we'll have to know whether or not Pg is deleting the underlying relation_filepath before and after the truncate command and if anything has that file open (if it exists). You may have to run lsof before the truncate too.

Related Solutions

PostgreSQL – Disk Space Not Released After TRUNCATE

According to comments in the source, truncate creates a new, empty storage file, and deletes the old storage file at commit time. (Docs suggest "storage file" is just a file as far as the OS is concerned, but I might be misunderstanding the terminology.)

Create a new empty storage file for the relation, and assign it as the relfilenode value. The old storage file is scheduled for deletion at commit.

Since it seems to be deleting a file, I can imagine some cases in which the underlying operating system might not immediately free that space. I imagine that in some cases the storage file might end up in the Recycling Bin under Windows, for example. But in my case, truncating a table under PostgreSQL 9.something immediately increased the freespace under Windows.

Truncation is also recorded in the WAL log. I don't know how much effect that might have.

Postgresql – Postgres multiple joins slow query, how to store default child record

You write:

Each customer can have multiple sites, but only one should be displayed in this list.

Yet, your query retrieves all rows. That would be a point to optimize. But you also do not define which site is to be picked.

Either way, it does not matter much here. Your EXPLAIN shows only 5026 rows for the site scan (5018 for the customer scan). So hardly any customer actually has more than one site. Did you ANALYZE your tables before running EXPLAIN?

From the numbers I see in your EXPLAIN, indexes will give you nothing for this query. Sequential table scans will be the fastest possible way. Half a second is rather slow for 5000 rows, though. Maybe your database needs some general performance tuning?

Maybe the query itself is faster, but "half a second" includes network transfer? EXPLAIN ANALYZE would tell us more.

If this query is your bottleneck, I would suggest you implement a materialized view.

After you provided more information I find that my diagnosis pretty much holds.

The query itself needs 27 ms. Not much of a problem there. "Half a second" was the kind of misunderstanding I had suspected. The slow part is the network transfer (plus ssh encoding / decoding, possibly rendering). You should only retrieve 100 rows, that would solve most of it, even if it means to execute the whole query every time.

If you go the route with a materialized view like I proposed you could add a serial number without gaps to the table plus index on it - by adding a column row_number() OVER (<your sort citeria here>) AS mv_id.

Then you can query:

SELECT *
FROM   materialized_view
WHERE  mv_id >= 2700
AND    mv_id <  2800;

This will perform very fast. LIMIT / OFFSET cannot compete, that needs to compute the whole table before it can sort and pick 100 rows.

pgAdmin timing

When you execute a query from the query tool, the message pane shows something like:

Total query runtime: 62 ms.

And the status line shows the same time. I quote pgAdmin help about that:

The status line will show how long the last query took to complete. If a dataset was returned, not only the elapsed time for server execution is displayed, but also the time to retrieve the data from the server to the Data Output page.

If you want to see the time on the server you need to use SQL EXPLAIN ANALYZE or the built in Shift + F7keyboard shortcut or Query -> Explain analyze. Then, at the bottom of the explain output you get something like this:

Total runtime: 0.269 ms

Best Answer

Related Solutions

PostgreSQL – Disk Space Not Released After TRUNCATE

Postgresql – Postgres multiple joins slow query, how to store default child record

pgAdmin timing

Related Question