Re: 8.2 Autovacuum BUG ?

[Pavan Teja <pavan.postgresdba@xxxxxxxxx>]

On Jan 24, 2018 7:57 PM, "Claudio Freire" <klaussfreire@xxxxxxxxx> wrote:

On Wed, Jan 24, 2018 at 8:50 AM, pavan95 <pavan.postgresdba@xxxxxxxxx> wrote:

Hello all,

One more interesting observation made by me.

I have ran the below query(s) on production:

SELECT
    relname,
    age(relfrozenxid) as xid_age,
    pg_size_pretty(pg_table_size(oid)) as table_size
FROM pg_class
WHERE relkind = 'r' and pg_table_size(oid) > 1073741824
ORDER BY age(relfrozenxid) DESC ;
                    relname                                              |
xid_age | table_size
------------------------------------------------------------+---------+------------
 *hxxxxxxxxxx*                                                      |
7798262 | 3245 MB
 hrxxxxxxxxx                                                         |
7797554 | 4917 MB
 irxxxxxxxxxx                                                        |
7796771 | 2841 MB
 hr_xxxxxxxxxxxxxxxx                                           | 7744262 |
4778 MB
 reimbxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx | 6767712 | 1110 MB

show autovacuum_freeze_max_age;
 autovacuum_freeze_max_age
---------------------------
 200000000
(1 row)

You seem to be rather far from the freeze_max_age. Unless you're consuming txids at a very high rate, I don't think that's your problem.

   Hi ,

     Yes, but why doing vacuum freeze of a table is causing a rapid ​archiving?? 

Any idea??

Regards,

Pavan