pg_repack needs to take an exclusive lock at the end of the
reorganization. If the lock cannot be taken after duration
--wait-timeout option specified and this option is true,
pg_repack gives up to repack a target table instead of
cancelling conflicting backend. False by default.
From inspecting the call sites of utoa, it appears that some of them
(especially the recent cleanup patch which added the strdup there) wanted
to prevent overwriting a local variable by repeated call (to utoa) using
the same variable as argument. This commit instead makes such call sites
strdup the variable itself before passing it to utoa. That seems cleaner
considering that it does not seem utoa's contract to do so (strdup its
parameter that is).
If there are two concurrent pg_repack commands are run on the same
table, the one starting later fails with error message:
Another pg_repack command may be running on the table. Please try again.
The document says this is shown as ERROR, but actualy is WARNING.
This patch includes one global counter (temp_obj_num) which counts number of temporary objects created by pg_repack. Correct order of deletion of temporary object as per count avoids unintentional error messages.
pgut version renamed to avoid confusion with the server version.
(I wonder why there is such a duplication of interfaces and
implementations there though...)
Previously, pg_repack shows "ERROR: ERROR: relation foo does not
exist" when specify non-existing table. Though the first ERROR
is from pg_repack and the second ERROR is from PostgreSQL server,
some users might think that pg_repack shows error level twice
wrongly.
"Waiting for %d transactions to finish. First PID: %s"
message. Display it on every loop through the SQL_XID_ALIVE check
(i.e. every second), instead of only when the number of transactions
we're waiting on changes -- previously, it was too easy for that
important message to get lost in other messages.
And don't display the message at all when running under pg_regress,
i.e. as part of `make installcheck`. We had been getting occasional
errors from pg_regress when autovacuum was running and that message
got logged.
These calls can require an access share lock on the table, which might
conflict with an existing or later acquires lock. So perform these calls
while we already have an exclusive lock on the table. This unfortuantely
means that we ave to remove the constness of the table parameter to
repack_one_table, as it is not modifying the table object to set up the
indexes.
problem was really about the OID being interpreted as an integer
literal upon input, and overflowing its integer space before even making
it into pg_try_advisory_lock(). (We do still need to add -2147483648 to
make the result fit into an integer, as 4b3347 does.)
Hopefully fixes issue #30, for real this time.
4-byte int accepted by the two-argument form of pg_try_advisory_lock()
we are using.
Fixes#30. Thanks to Mark Steben and Greg Sabino Mullane for the report
and diagnosis.
Adds support for repacking only the tables in a specified schema. This
doesn't support --only-indexes mode, but that seems alright for now.
Fix merge conflicts, and make a few tweaks along the way:
* bump version to 1.3-dev0
* add Beena to list of maintainers
* documentation wordsmithing
* fix up the INFO message printed for each index in --index or
--only-indexes mode, so that it is only printed once per index, and
prints the name of the original index, not that of the transient
index_%u name.
It was possible to specify both --schema and --table which probably
-should- be legal but would need some code to be rewritten. This patch
adds a check that both can't be specified and returns an error telling
the user to use schema.table notation instead. A regression test
checking this behaviour was added.
problem: in case there are open transactions on other databases then the
one pg_repack is working on and pg_locks doesn't contain any
information about the affected database oid of the locked relation
(e.g. there is no locked relation, only an open transaction),
pg_repack will wait for that connection to release the lock
(even if the relation that gets reorganized is held in an
different database).
solution: join pg_database (via pg_stat_activity's datid) and check
if the connection (of the conflicted transaction) is established
on a different database than the relation treated by pg_repack
and skip them.
furthermore don't exclude transactions from other databases when
shared objects are locked.
* More mentions of new --only-indexes feature
* Note we now support up to Postgres 9.3, and get rid of outdated list
of supported operating systems. (As far as we know, pg_repack should
build on any platforms supported by PostgreSQL itself, although no one
has tested the Windows build in a long time.)
* Remove most of the warnings about data corruption possible with concurrent
DDL: this should no longer be a concern now that we hold an ACCESS SHARE
lock during full-table repacks. Keep a short warning about old versions
being susceptible to this problem, just to make clear that it's fixed now
and as an enticement to upgrade.
* A few grammar, phrasing, and typo fixes
Previously, any error creating the temp index was blamed on a previous
temporary index existing. However, other errors could occur, e.g. if
--tablespace=pg_global was specified. Perform an explicit check for
whether an existing index_xxxxx already exists first, so that we
can give that error message only in the appropriate case.
* Avoid using create_idx from a PGresult that's already been PQclear'ed
* Now that we're properly tracking which indexes have been repacked,
don't bother with IF EXISTS in DROP INDEX. Better to make any errors
here stand out.
* Fix error handling in repack_table_indexes() to keep track of which
indexes were actually rebuilt successfully by pg_repack, so that only
those indexes will be dropped. Previously, we would issue an error
message telling the user to use DROP INDEX but then just drop the
index anyway.
* DROP INDEX command needs to quote the schema name in case it is
not simple.
* missing CLEARPGRES() in repack_all_indexes()
Improve locking behavior in indexes-only mode so we only need one
ACCESS EXCLUSIVE lock per table instead of one per index. Support
multiple --index options. Fix for schema-name parsing.
Patch from Beena Emerson.
Switch to using the two-input form of pg_advisory_lock(), so as
to avoid impacting other applications which might happen to lock
just the OID of the table. The REPACK_LOCK_PREFIX_STR is a decimal
version of the first three bytes of
echo -n "pg_repack" | sha1sum
