Re: Why will vacuum not end?

From: "Shea,Dan [CIS]" <Dan(dot)Shea(at)ec(dot)gc(dot)ca>
To: 'Josh Berkus' <josh(at)agliodbs(dot)com>
Cc: pgsql-performance(at)postgresql(dot)org
Subject: Re: Why will vacuum not end?
Date: 2004-04-24 19:48:19
Message-ID: F2D63B916C88C14D9B59F93C2A5DD33F0B9252@cisxa.cmc.int.ec.gc.ca
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

Manfred is indicating the reason it is taking so long is due to the number
of dead tuples in my index and the vacuum_mem setting.
The last delete that I did before starting a vacuum had 219,177,133
deletions.
Dan.
>Dan,

>> Josh, how long should a vacuum take on a 87 GB table with a 39 GB index?

>Depends:
>-- What's your disk support?

>-- VACUUM, VACUUM ANALYZE, or VACUUM FULL?
VACUUM ANALYZE
>-- What's your vacuum_mem setting?
set vacuum_mem = '196608'
#fsync = true # turns forced synchronization on or off
#wal_sync_method = fsync
>-- What are checkpoint and wal settings?
wal_buffers = 64
checkpoint_segments = 30
checkpoint_timeout = 300

>> I see alot of activity with transaction logs being recycled (15 to 30
every
>> 3 to 20 minutes).
>> Is the vacuum causing this?

>Probably, yes. How many checkpoint_buffers do you allow?
I am not sure what the checkpoint_buffers are, we are running 7.4.0?
>--
>Josh Berkus
>Aglio Database Solutions
>San Francisco

Responses

Browse pgsql-performance by date

  From Date Subject
Next Message Shea,Dan [CIS] 2004-04-24 19:58:08 Re: Why will vacuum not end?
Previous Message Manfred Koizar 2004-04-24 17:57:22 Re: Why will vacuum not end?