From: | Andrew Gierth <andrew(at)tao11(dot)riddles(dot)org(dot)uk> |
---|---|
To: | Horst Düster <horst(dot)duester(at)sourcepole(dot)ch> |
Cc: | pgsql-admin(at)lists(dot)postgresql(dot)org |
Subject: | Re: Huge table pg_shdepend |
Date: | 2019-01-21 11:19:58 |
Message-ID: | 87o98a468n.fsf@news-spur.riddles.org.uk |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-admin |
>>>>> "Horst" == Horst Düster <horst(dot)duester(at)sourcepole(dot)ch> writes:
Horst> Hi
Horst> I'm struggling with a huge table pg_shdepend. My database has a
Horst> size of 640MB exclude the size of the system tables. But the
Horst> real size of the db is more than 2GB. The reason is a huge table
Horst> pg_shdepend with more than 32000000 records.
pg_shdepend has a row for every object and every ACL entry that records
which role (other than the initial db superuser) owns the object or is
referenced from the ACL entry.
Do you have an unusually large number of tables (or other database
objects)?
Note that there's a single pg_shdepend table shared between all
databases in the cluster; it's not part of any single database but
rather is a "global" table. So what matters is the total number of
objects over all databases, not the size of any single db.
--
Andrew (irc:RhodiumToad)
From | Date | Subject | |
---|---|---|---|
Next Message | Ron | 2019-01-21 11:42:02 | Re: Huge table pg_shdepend |
Previous Message | AB_ba# | 2019-01-21 11:04:46 | IO scheduler recommendation |