Re: [HACKERS] Press Release Draft - 2016-02-09 Cumulative Update

On Feb 7, 2017, at 12:44 PM, Alvaro Herrera <alvherre@2ndquadrant.com> wrote:

Jonathan S. Katz wrote:

Below is the draft of the press release for the update this Thursday:

https://git.postgresql.org/gitweb/?p=press.git;a=blob;f=update_releases/current/20170209updaterelease.md;h=0cccb8986c08527f65f13d704a78c36bb8de7fef;hb=afc01091dea8a1597e8e21430edc3908c581ce0c <https://git.postgresql.org/gitweb/?p=press.git;a=blob;f=update_releases/current/20170209updaterelease.md;h=0cccb8986c08527f65f13d704a78c36bb8de7fef;hb=afc01091dea8a1597e8e21430edc3908c581ce0c>

As there are a lot of updates I did my best to consolidate some of the bullet points and as usual, people are directed to the release notes.  Please let me know if there are any inaccuracies so I can fix them ASAP.

Please do post the proposed text on list for ease of review.  I wasn't
looking at the text, so I wouldn't have noticed this if Emre hadn't
replied:

 76 If you believe you have been affected by the aforementioned CREATE INDEX CONCURRENTLY bug, you will have to rebuild the index.  An example of rebuilding an index:
 77
 78     BEGIN;
 79     DROP INDEX bad_index_name;
 80     CREATE INDEX CONCURRENTLY bad_index_name ON table_name (column_name); /* replace names with your original index definition */
 81     COMMIT;

This is not a good recipe, because using CREATE INDEX CONCURRENTLY in
the same transaction that grabs an exclusive lock on the table for the
DROP INDEX is pointless -- the access exclusive lock is held until the
end of the transaction, and CIC does not work inside a transaction
anyway so it'd raise an ERROR and rollback the DROP INDEX.  So the user
would probably drop the BEGIN/COMMIT sequence in order for this to work
in the first place.  (The other option is to use CREATE INDEX not
concurrent, but that would lock the table for a very long time).