Need help understanding vacuum verbose output

From: Gordon Shannon <gordo169(at)gmail(dot)com>
To: pgsql-general(at)postgresql(dot)org
Subject: Need help understanding vacuum verbose output
Date: 2010-08-05 20:00:50
Message-ID: 1281038450504-2265895.post@n5.nabble.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general


Hi, Running 8.4.4 on Centos. A couple of these numbers don't make sense to
me.
(I added line numbers for reference)

1 vacuum verbose authors_archive;
2 INFO: vacuuming "public.authors_archive"
3 INFO: scanned index "authors_archive_pkey" to remove 45878 row versions
4 DETAIL: CPU 0.05s/0.34u sec elapsed 0.41 sec.
5 INFO: "authors_archive": removed 45878 row versions in 396 pages
6 DETAIL: CPU 0.00s/0.00u sec elapsed 0.00 sec.
7 INFO: index "authors_archive_pkey" now contains 5012898 row versions in
13886 pages
8 DETAIL: 45878 index row versions were removed.
9 138 index pages have been deleted, 12 are currently reusable.
10 CPU 0.00s/0.00u sec elapsed 0.00 sec.
11 INFO: "authors_archive": found 16558 removable, 7300 nonremovable row
versions in 492 out of 51958 pages
12 DETAIL: 0 dead row versions cannot be removed yet.
13 There were 4687 unused item pointers.
14 0 pages are entirely empty.

I was the only one touching this table. Since the previous vacuum, I deleted
45,878 rows.
Line 5 says it removed all 45,878 tuples, that's fine.
Line 11 now says it found only 16,558 removable tuples. What does this mean?
Line 11 also says there are 7,300 non-removable tuples. I expected 0. What
does this mean?
Line 12 says that 0 dead tuples cannot be removed, which seems correct to
me, but seems to contradict Line 11.

Thanks.

--
View this message in context: http://postgresql.1045698.n5.nabble.com/Need-help-understanding-vacuum-verbose-output-tp2265895p2265895.html
Sent from the PostgreSQL - general mailing list archive at Nabble.com.

Responses

Browse pgsql-general by date

  From Date Subject
Next Message Edmundo Robles L. 2010-08-05 21:37:17 Relation raltion_name with OID ##### no longer exists
Previous Message Greg Smith 2010-08-05 19:43:58 Re: Why create CHECKPOINT_WAIT checkpoint in pg_start_backup not CHECKPOINT_IMMEDIATE?