Re: pgsql: Fix BRIN to use SnapshotAny during summarization

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org>, pgsql-committers(at)postgresql(dot)org
Subject: Re: pgsql: Fix BRIN to use SnapshotAny during summarization
Date: 2015-08-07 03:22:39
Message-ID: 19440.1438917759@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-committers pgsql-hackers

... btw, I believe the reason for the failure here:
http://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=jaguarundi&dt=2015-08-06%2011%3A52%3A17

is that brin_page_items() is unsafe. It's accessing
state->bdesc->bd_tupdesc, which is pointing at the tupdesc of an index
Relation that it no longer has open; not only directly, but via
brin_deform_tuple(). All you need is a relcache flush to be accessing
garbage. I haven't dug down thoroughly, but I'd not be surprised if there
were also some dereferences of bd_index, which is equally a dangling
pointer.

regards, tom lane

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Noah Misch 2015-08-07 03:23:43 Re: Raising our compiler requirements for 9.6
Previous Message Merlin Moncure 2015-08-07 03:04:42 Re: Autonomous Transaction is back

Browse pgsql-committers by date

  From Date Subject
Next Message Andres Freund 2015-08-07 13:13:39 pgsql: Address points made in post-commit review of replication origins
Previous Message Bruce Momjian 2015-08-07 02:34:08 pgsql: 9.5 release notes: updates from Andres Freund and Jeff Janes