Обсуждение: vacuum full doubled database size
We did a "vacuum full" on a database which had been interrupted by a network outage.
We found the database size doubled afterwards.
Autovacuum also found a lot of orphaned tables afterwards.
The ophan temp objects went away after a cluster restart while the db size remained doubled.
Any idea?
Postgres 9.6.17
We did a "vacuum full" on a database which had been interrupted by a network outage.
We found the database size doubled afterwards.
Autovacuum also found a lot of orphaned tables afterwards.
The ophan temp objects went away after a cluster restart while the db size remained doubled.
Any idea?
Postgres 9.6.17
"Genius might be described as a supreme capacity for getting its possessors
into trouble of all kinds."
-- Samuel Butler
Thanks. How to get rid of it. New vacuum full?
Von: Michael Loftis <mloftis@wgops.com>
Gesendet: Freitag, 13. März 2020 14:48
An: Zwettler Markus (OIZ) <Markus.Zwettler@zuerich.ch>
Cc: pgsql-general <pgsql-general@postgresql.org>
Betreff: Re: vacuum full doubled database size
A vacuum full rebuilds the tables, so yeah if it didn’t successfully complete I would expect a lot of dead data.
On Fri, Mar 13, 2020 at 07:41 Zwettler Markus (OIZ) <Markus.Zwettler@zuerich.ch> wrote:
We did a "vacuum full" on a database which had been interrupted by a network outage.
We found the database size doubled afterwards.
Autovacuum also found a lot of orphaned tables afterwards.
The ophan temp objects went away after a cluster restart while the db size remained doubled.
Any idea?
Postgres 9.6.17
--
"Genius might be described as a supreme capacity for getting its possessors
into trouble of all kinds."
-- Samuel Butler
We did a "vacuum full" on a database which had been interrupted by a network outage.
We found the database size doubled afterwards.
Autovacuum also found a lot of orphaned tables afterwards.
The ophan temp objects went away after a cluster restart while the db size remained doubled.
Any idea?
Postgres 9.6.17
Angular momentum makes the world go 'round.
On 13/03/2020 15:15, Ron wrote: > This is why I'd VACUUM FULL in a planned manner, one or two tables at a > time, and *locally* from crontab. That's not really viable on any remotely busy system: VACUUM FULL claims exclusive table locks, causing queries to hang (https://www.postgresql.org/docs/current/sql-vacuum.html#NOTES mentions this too). Tools like pg_repack can do some live shrinking. I've also had some success at reclaiming space without large scale locks by carefully crafting some atomic DELETE + INSERT in order to force tuples from the tail end into gaps at lower CTIDs (physical page address) that were made available by previous plain VACUUMs - regular VACUUM will clip data files if all the tuples beyond a given offset are dead. > > On 3/13/20 8:41 AM, Zwettler Markus (OIZ) wrote: >> >> We did a "vacuum full" on a database which had been interrupted by a >> network outage. >> >> We found the database size doubled afterwards. >> >> Autovacuum also found a lot of orphaned tables afterwards. >> >> The ophan temp objects went away after a cluster restart while the db >> size remained doubled. >> >> Any idea? >> >> Postgres 9.6.17 >> > > -- > Angular momentum makes the world go 'round. -- Regards Fabio Ugo Venchiarutti OSPCFC Network Engineering Dpt. Ocado Technology -- Notice: This email is confidential and may contain copyright material of members of the Ocado Group. Opinions and views expressed in this message may not necessarily reflect the opinions and views of the members of the Ocado Group. If you are not the intended recipient, please notify us immediately and delete all copies of this message. Please note that it is your responsibility to scan this message for viruses. References to the "Ocado Group" are to Ocado Group plc (registered in England and Wales with number 7098618) and its subsidiary undertakings (as that expression is defined in the Companies Act 2006) from time to time. The registered office of Ocado Group plc is Buildings One & Two, Trident Place, Mosquito Way, Hatfield, Hertfordshire, AL10 9UL.
>Fabio Ugo Venchiarutti wrote: >On 13/03/2020 15:15, Ron wrote: >> This is why I'd VACUUM FULL in a planned manner, one or two tables at >> a time, and *locally* from crontab. > >That's not really viable on any remotely busy system: VACUUM FULL claims exclusive table locks, causing queries to hang(https://www.postgresql.org/docs/current/sql-vacuum.html#NOTES mentions this too). > >Tools like pg_repack can do some live shrinking. To say "not really viable on any remotely busy system" is a pretty sweeping statement. I think a better statement is that "for many busy systems, this could be a real problem and to consider it carefully in light of your needs." On our systems, we do this just fine, though the difference is probably the level of busy. We have periods that are not as busy as others (1am). In addition, most of our tables are fairly small'ish and a VACUUM FULL takes 30sec or less, so it's not so bad. The vast majority of our data is in about a dozen tables which are mostly used for reports by people in the daytime, so if they lock for 5min each in the middle of the night on a weekend it's OK (and we only do this once a quarter). So this approach can work, but "it depends" and "YMMV" and all that jazz...which I believe was what Ron was trying to point out with planning. That being said, I've had "go check out pg_repack" on my to-do list for a while and one day I will, but at the moment the above works for us. Kevin . This e-mail transmission, and any documents, files or previous e-mail messages attached to it, may contain confidential information.If you are not the intended recipient, or a person responsible for delivering it to the intended recipient, youare hereby notified that any disclosure, distribution, review, copy or use of any of the information contained in or attachedto this message is STRICTLY PROHIBITED. If you have received this transmission in error, please immediately notifyus by reply e-mail, and destroy the original transmission and its attachments without reading them or saving them todisk. Thank you.
On 13/03/2020 15:15, Ron wrote:This is why I'd VACUUM FULL in a planned manner, one or two tables at a time, and *locally* from crontab.
That's not really viable on any remotely busy system: VACUUM FULL claims exclusive table locks, causing queries to hang (https://www.postgresql.org/docs/current/sql-vacuum.html#NOTES mentions this too).
This is of course if I were to ever do a VACUUM FULL... :)
The point is that your mother's rule about not sticking a giant fork-full of food into your mouth is also an excellent rule in IT.
--