Re: BUG #16045: vacuum_db crash and illegal memory alloc after pg_upgrade from PG11 to PG12

Bruce Momjian <bruce@momjian.us>

From: Bruce Momjian <bruce@momjian.us>

To: Tomas Vondra <tomas.vondra@2ndquadrant.com>

Cc: Tom Lane <tgl@sss.pgh.pa.us>, Andres Freund <andres@anarazel.de>, buschmann@nidsa.net, pgsql-bugs@lists.postgresql.org

Date: 2019-10-23T22:08:21Z

Lists: pgsql-bugs, pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

Move into separate file all the SQL queries used in pg_upgrade tests
- 1924d508c335 10.20 landed
- 0e603b75c434 11.15 landed
- b6e525648d72 12.10 landed
- fae5f08e1719 13.6 landed
- b6dac98b0561 14.2 landed
- 0df9641d3905 15.0 landed
Add table to regression tests for binary-compatibility checks in pg_upgrade
- a9993416f80f 12.10 landed
- 755f04c72ef1 13.6 landed
- cf3d79aa31f2 14.2 landed
- 835bcba8b8d7 15.0 landed
Fix tests of pg_upgrade across different major versions
- afa09e4a9af6 12.9 landed
- 2a8dee6a67cc 13.5 landed
- f4e1c8892b9e 14.1 landed
- fa66b6dee084 15.0 landed
Multirange datatypes
- 6df7a9698bb0 14.0 cited
Work around cross-version-upgrade issues created by commit 9e38c2bb5.
- 97f73a978fc1 14.0 cited
Declare assorted array functions using anycompatible not anyelement.
- 9e38c2bb5093 14.0 cited
Remove factorial operators, leaving only the factorial() function.
- 76f412ab3105 14.0 cited
Create by default sql/ and expected/ for output directory in pg_regress
- e78900afd217 14.0 cited
Add missing include to pg_upgrade/version.c
- bc3a94dc0005 9.4.25 landed
- 984aa0ede1d2 9.5.20 landed
- e09ab32a2205 9.6.16 landed
Improve the check for pg_catalog.line data type in pg_upgrade
- 235a52ca0f26 9.4.25 landed
- f57b01dd75ee 9.5.20 landed
- 0a643de08715 9.6.16 landed
- 2218fdca496b 10.11 landed
- a970b6cdebd1 11.6 landed
- ebb4caa9120d 12.1 landed
- 8d48e6a7240c 13.0 landed
Improve the check for pg_catalog.unknown data type in pg_upgrade
- e86ece22114d 10.11 landed
- d071a2539ff4 11.6 landed
- a8e49ae0c381 12.1 landed
- a524f50d0fc6 13.0 landed
Check for tables with sql_identifier during pg_upgrade
- eaf900e842ab 12.1 landed
- 0ccfc2822366 13.0 landed
pg_upgrade: clarify the database names in error files
- 1634d361577a 13.0 cited
In the pg_upgrade test suite, don't write to src/test/regress.
- 40b132c1afbb 12.0 cited
Allow group access on PGDATA
- c37b3d08ca68 11.0 cited
Refactor dir/file permissions
- da9b580d8990 11.0 cited
Remove unused functions in regress.c.
- db3af9feb19f 11.0 cited
Make WAL segment size configurable at initdb time.
- fc49e24fa69a 11.0 cited
Fix bit-rot in pg_upgrade's test.sh, and improve documentation.
- 5bab1985dfc2 10.0 cited

Attachments

infoschema.diff (text/x-diff) patch

On Tue, Oct 15, 2019 at 02:18:17AM +0200, Tomas Vondra wrote:
> On Mon, Oct 14, 2019 at 06:35:38PM +0200, Tomas Vondra wrote:
> > ...
> > 
> > Aha! I forgot we copy the necessary stuff into pg_attribute. Thanks for
> > clarifying, I'll polish and push the fix shortly.
> > 
> 
> I've pushed and backpatched the fix. Attached are similar fixes for the
> existing pg_upgrade checks for pg_catalog.line and pg_catalog.unknown
> types, which have the same issues with composite types and domains.

This comit added old_11_check_for_sql_identifier_data_type_usage(), but
it did not use the clearer database error list format added to the
master branch in commit 1634d36157.  Attached is a patch to fix this,
which I have committed.

-- 
  Bruce Momjian  <bruce@momjian.us>        http://momjian.us
  EnterpriseDB                             http://enterprisedb.com

+ As you are, so once was I.  As I am, so you will be. +
+                      Ancient Roman grave inscription +