Thank you Igor, I was able to eliminate the 15GB bloating for a 35GB
table size , only after I restart the Pg server with one single
connections and run a full vacuum for table.
Isabella
On 10/01/18 11:10 AM, Igor Neyman wrote:
> -----Original Message-----
> From: Isabella Ghiurea [mailto:isabella.ghiurea@nrc-cnrc.gc.ca]
> Sent: Wednesday, January 10, 2018 10:48 AM
> To: pgsql-performance@postgresql.org
> Subject: RE: PG 9.5 2 tables same DDL with diff size
>
> Attention: This email was sent from someone outside of Perceptron. Always exercise caution when opening attachments
orclicking links from unknown senders or when receiving unexpected emails.
>
>
> I run full vacuum and reindex on largest table (50GB) while there was no server activities so I assume no
transactionwas holding a lock on table since the full vacuum was able to run, anything where I should consider looking
?
>
>
> __________________________________________________________________________________________________________
>
> Yes, in pg_stat_activity look for idle transactions that started long time ago.
> To prevent vacuum from doing its job they don't need to lock the table, they could just prevent from cleaning "old"
rowversions.
>
> Regards,
> Igor Neyman
>