New bug fix releases: 8.4.8, 8.3.15

[Martin Pitt]

Binary package hint: postgresql-8.4

PostgreSQL published new microreleases which fix a major potential data-loss bug with using pg_upgrade to 9.0, and some other fixes. No security updates. Please see http://www.postgresql.org/about/news.1310 for details.

As per the standing microrelease exception these should go into stables.

[Martin Pitt (pitti)]

Note that 8.1 is discontinued upstream, so we won't get a dapper update. Karmic is about to go EOL, so we won't update that either any more.

[Martin Pitt (pitti)]

Question is whether this should go into natty final (by way of syncing from sid), or wait for an SRU? Personally I'd favor doing a sync, as it's well covered by test suites, but I don't want to make the -release call in this case.

[Scott Kitterman (kitterman)]

I second doing it now.

[Martin Pitt (pitti)]

Maverick update uploaded, waiting in unapproved queue.

[Martin Pitt (pitti)]

Oh, Natty will be an upload, not a sync, as in sid I dropped the client-side libraries there (9.0 is the default in sid now).

[Martin Pitt (pitti)]

Lucid update uploaded to queue for review, too.

[Clint Byrum (clint-fewbar)]

Accepted postgresql-8.4 into maverick-proposed, the package will build now and be available in a few hours. Please test and give feedback here. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you in advance!

[Martin Pitt (pitti)]

Natty update uploaded as well, waiting in unapproved.

[Martin Pitt (pitti)]

Hardy update uploaded to the queue for review as well.

[Clint Byrum (clint-fewbar)]

Accepted postgresql-8.4 into lucid-proposed, the package will build now and be available in a few hours. Please test and give feedback here. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you in advance!

[Clint Byrum (clint-fewbar)]

Accepted postgresql-8.3 into hardy-proposed, the package will build now and be available in a few hours. Please test and give feedback here. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you in advance!

[Launchpad Janitor (janitor)]

This bug was fixed in the package postgresql-8.4 - 8.4.8-0ubuntu0.11.04

---------------
postgresql-8.4 (8.4.8-0ubuntu0.11.04) natty; urgency=low

  * New upstream bug fix release: (LP: #767165)
    - If your installation was upgraded from a previous major release by
      running pg_upgrade, you should take action to prevent possible data loss
      due to a now-fixed bug in pg_upgrade. The recommended solution is to run
      "VACUUM FREEZE" on all TOAST tables. More information is available at
      http://wiki.postgresql.org/wiki/20110408pg_upgrade_fix.
    - Fix pg_upgrade's handling of TOAST tables.
      This error poses a significant risk of data loss for installations
      that have been upgraded with pg_upgrade. This patch corrects the
      problem for future uses of pg_upgrade, but does not in itself cure
      the issue in installations that have been processed with a buggy
      version of pg_upgrade.
    - Suppress incorrect "PD_ALL_VISIBLE flag was incorrectly set"
      warning.
    - Disallow including a composite type in itself.
    - Avoid potential deadlock during catalog cache initialization.
    - Fix dangling-pointer problem in BEFORE ROW UPDATE trigger handling
      when there was a concurrent update to the target tuple.
    - Disallow "DROP TABLE" when there are pending deferred trigger
      events for the table.
      Formerly the "DROP" would go through, leading to "could not open
      relation with OID nnn" errors when the triggers were eventually
      fired.
    - Prevent crash triggered by constant-false WHERE conditions during
      GEQO optimization.
    - Improve planner's handling of semi-join and anti-join cases.
    - Fix selectivity estimation for text search to account for NULLs.
    - Improve PL/pgSQL's ability to handle row types with dropped columns.
    - Fix PL/Python memory leak involving array slices.
    - Fix pg_restore to cope with long lines (over 1KB) in TOC files.
    - Put in more safeguards against crashing due to division-by-zero
      with overly enthusiastic compiler optimization. (Closes: #616180)
 -- Martin Pitt <email address hidden> Wed, 20 Apr 2011 16:10:45 +0200

[Martin Pitt (pitti)]

All versions passed the upstream test suite on all architectures, as otherwise they would have failed to build. There have been no regression reports on the upstream bug lists.

I have run the postgresql-common integration test on hardy, lucid, and maverick, and confirm that they all pass.

[Launchpad Janitor (janitor)]

This bug was fixed in the package postgresql-8.3 - 8.3.15-0ubuntu8.04

---------------
postgresql-8.3 (8.3.15-0ubuntu8.04) hardy-proposed; urgency=low

  * New upstream bug fix release: (LP: #767165)
    - Disallow including a composite type in itself.
    - Avoid potential deadlock during catalog cache initialization.
    - Fix dangling-pointer problem in BEFORE ROW UPDATE trigger handling
      when there was a concurrent update to the target tuple.
    - Disallow "DROP TABLE" when there are pending deferred trigger
      events for the table.
    - Fix PL/Python memory leak involving array slices.
    - Fix pg_restore to cope with long lines (over 1KB) in TOC files.
    - Put in more safeguards against crashing due to division-by-zero
      with overly enthusiastic compiler optimization.
 -- Martin Pitt <email address hidden> Wed, 20 Apr 2011 16:20:32 +0200

[Launchpad Janitor (janitor)]

This bug was fixed in the package postgresql-8.4 - 8.4.8-0ubuntu0.10.04

---------------
postgresql-8.4 (8.4.8-0ubuntu0.10.04) lucid-proposed; urgency=low

  * New upstream bug fix release: (LP: #767165)
    - If your installation was upgraded from a previous major release by
      running pg_upgrade, you should take action to prevent possible data loss
      due to a now-fixed bug in pg_upgrade. The recommended solution is to run
      "VACUUM FREEZE" on all TOAST tables. More information is available at
      http://wiki.postgresql.org/wiki/20110408pg_upgrade_fix.
    - Fix pg_upgrade's handling of TOAST tables.
      This error poses a significant risk of data loss for installations
      that have been upgraded with pg_upgrade. This patch corrects the
      problem for future uses of pg_upgrade, but does not in itself cure
      the issue in installations that have been processed with a buggy
      version of pg_upgrade.
    - Suppress incorrect "PD_ALL_VISIBLE flag was incorrectly set"
      warning.
    - Disallow including a composite type in itself.
    - Avoid potential deadlock during catalog cache initialization.
    - Fix dangling-pointer problem in BEFORE ROW UPDATE trigger handling
      when there was a concurrent update to the target tuple.
    - Disallow "DROP TABLE" when there are pending deferred trigger
      events for the table.
      Formerly the "DROP" would go through, leading to "could not open
      relation with OID nnn" errors when the triggers were eventually
      fired.
    - Prevent crash triggered by constant-false WHERE conditions during
      GEQO optimization.
    - Improve planner's handling of semi-join and anti-join cases.
    - Fix selectivity estimation for text search to account for NULLs.
    - Improve PL/pgSQL's ability to handle row types with dropped columns.
    - Fix PL/Python memory leak involving array slices.
    - Fix pg_restore to cope with long lines (over 1KB) in TOC files.
    - Put in more safeguards against crashing due to division-by-zero
      with overly enthusiastic compiler optimization. (Closes: #616180)
 -- Martin Pitt <email address hidden> Wed, 20 Apr 2011 15:57:49 +0200

[Launchpad Janitor (janitor)]

This bug was fixed in the package postgresql-8.4 - 8.4.8-0ubuntu0.10.10

---------------
postgresql-8.4 (8.4.8-0ubuntu0.10.10) maverick-proposed; urgency=low

  * New upstream bug fix release: (LP: #767165)
    - If your installation was upgraded from a previous major release by
      running pg_upgrade, you should take action to prevent possible data loss
      due to a now-fixed bug in pg_upgrade. The recommended solution is to run
      "VACUUM FREEZE" on all TOAST tables. More information is available at
      http://wiki.postgresql.org/wiki/20110408pg_upgrade_fix.
    - Fix pg_upgrade's handling of TOAST tables.
      This error poses a significant risk of data loss for installations
      that have been upgraded with pg_upgrade. This patch corrects the
      problem for future uses of pg_upgrade, but does not in itself cure
      the issue in installations that have been processed with a buggy
      version of pg_upgrade.
    - Suppress incorrect "PD_ALL_VISIBLE flag was incorrectly set"
      warning.
    - Disallow including a composite type in itself.
    - Avoid potential deadlock during catalog cache initialization.
    - Fix dangling-pointer problem in BEFORE ROW UPDATE trigger handling
      when there was a concurrent update to the target tuple.
    - Disallow "DROP TABLE" when there are pending deferred trigger
      events for the table.
      Formerly the "DROP" would go through, leading to "could not open
      relation with OID nnn" errors when the triggers were eventually
      fired.
    - Prevent crash triggered by constant-false WHERE conditions during
      GEQO optimization.
    - Improve planner's handling of semi-join and anti-join cases.
    - Fix selectivity estimation for text search to account for NULLs.
    - Improve PL/pgSQL's ability to handle row types with dropped columns.
    - Fix PL/Python memory leak involving array slices.
    - Fix pg_restore to cope with long lines (over 1KB) in TOC files.
    - Put in more safeguards against crashing due to division-by-zero
      with overly enthusiastic compiler optimization. (Closes: #616180)
 -- Martin Pitt <email address hidden> Wed, 20 Apr 2011 15:33:25 +0200