- 13 1月, 2018 40 次提交
-
-
由 Heikki Linnakangas 提交于
On second thoughts, let's keep the QD mirror running. I'm not sure why we disabled it originally, but now that I went through the trouble of fixing subsequent tests to pass with a QD mirror, let's keep it enabled.
-
由 Heikki Linnakangas 提交于
-
由 Heikki Linnakangas 提交于
The setup script for the walrep tests disables the standby for the QD node. That used to not work, and therefore the mirror was still running, but now that I fixed "gpinitstandby -ra", it is disabled again. I'm not convinced that we should disable the mirror, and leave it disabled, in the walrep tests. But that's the way it was.
-
由 Heikki Linnakangas 提交于
-
由 Heikki Linnakangas 提交于
I had removed the fileSpaceDictionary argument to SegmentRow constructor earlier, but this was left behind.
-
由 Heikki Linnakangas 提交于
-
由 Heikki Linnakangas 提交于
print_summary() no longer has a filespaces argument.
-
由 Heikki Linnakangas 提交于
-
由 Heikki Linnakangas 提交于
-
由 Heikki Linnakangas 提交于
Persistent table are no more.
-
由 Heikki Linnakangas 提交于
We might want to have something similar with tablespaces, but not as a TINC test like this.
-
由 Heikki Linnakangas 提交于
-
由 Heikki Linnakangas 提交于
-
由 Heikki Linnakangas 提交于
These functions don't exist anymore.
-
由 Heikki Linnakangas 提交于
There are some "known" differences that are currently not masked out. I think the 'last_value' field is WAL-logged to a higher value than what's written to disk in the master. This was not visible before, when the demo cluster didn't have a standby for the QD node. In QE nodes, sequences are not used.
-
由 Heikki Linnakangas 提交于
-
由 Xiaoran Wang 提交于
Signed-off-by: NMax Yang <myang@pivotal.io>
-
由 Ashwin Agrawal 提交于
Also, remove caling GpRecover() in test_mpp23395. Ideally this test needs to be re-written in isolation2 but this should atleast get it green.
-
由 Heikki Linnakangas 提交于
Now that max_wal_senders is not pegged at 1, it's normal for there to be unused slots.
-
由 Jacob Champion 提交于
Follow-up to the previous commit.
-
由 Heikki Linnakangas 提交于
-
由 Heikki Linnakangas 提交于
It's now in gp_segment_configuration.
-
由 Heikki Linnakangas 提交于
-
由 Heikki Linnakangas 提交于
I removed some fields from here earlier, update unit test accordingly.
-
由 Heikki Linnakangas 提交于
-
由 Heikki Linnakangas 提交于
-
由 Heikki Linnakangas 提交于
-
由 Heikki Linnakangas 提交于
-
由 Heikki Linnakangas 提交于
The gpinitsystem -F option is now just path to the main data dir, without a filespace name.
-
由 Heikki Linnakangas 提交于
With ORCA, the CREATE TABLE AS used in the test created a randomly distributed table, while with the Postgres planner, it was distributed by the only column. A randomly distributed table cannot have an index, so you got an error with that. Fix by forcing the distribution.
-
由 Heikki Linnakangas 提交于
I'm actually not sure why the QD now suddenly has a standby in the default demo cluster. That seems to have happened as a side-effect of the filespace removal commit. That change was supposedly made already in commit 67c9ab91b9 already, on Dec 20th, but seems that it didn't actually happen until now, for some reason. In any case, that is what we want, so adjust the tests to cope.
-
由 Heikki Linnakangas 提交于
The test uses as a replication connection, so it didn't work with max_wal_senders=1 if the QD node has a standby. And it seems like a good idea to have some headroom, anyway. At some point, we ought to make this a true GUC like in the upstream, but this will do for now.
-
由 Heikki Linnakangas 提交于
There are more references in other tools, but we're not testing those yet.
-
由 Heikki Linnakangas 提交于
This is a backport upstream commit 22817041, from PostgreSQL 9.0, which added the server version number in the path. But in GPDB, also include the gp_dbid in the path. This makes it possible to use the same tablespace path on multiple servers running on the same host, without clashing. Also includes cherry-pick of the small upstream cleanup commits 5c82ccb1 a6f56efc, and c282b36d. Re-enable upstream 'tablespace' regression test. It works now, even when all the nodes are running on same host.
-
由 Heikki Linnakangas 提交于
Remove the concept of filespaces, revert tablespaces to work the same as in upstream. There is some leftovers in management tools. I don't know how to test all that, and I was afraid of touching things I can't run. Also, we may need to create replacements for some of those things on top of tablespaces, to make the management of tablespaces easier, and it might be easier to modify the existing tools than write them from scratch. (Yeah, you could always look at the git history, but still.) Per the discussion on gpdb-dev mailing list, the plan is to cherry-pick commit 16d8e594 from PostgreSQL 9.2, to make it possible to have a different path for a tablespace in the primary and its mirror. But that's not included in this commit yet. TODO: Make temp_tablespaces work. TODO: Make pg_dump do something sensible, when dumping from a GPDB 5 cluster that uses filespaces. Same with pg_upgrade. Discussion: https://groups.google.com/a/greenplum.org/d/msg/gpdb-dev/sON4lraPEqg/v3lkM587BAAJ
-
由 Heikki Linnakangas 提交于
-
由 Heikki Linnakangas 提交于
In order to debug the MM_gpexpand_2 test failure. TINC prints the full exception in the TINC log, I think, but I have no access to the log on the CI machine, and I have not been able to run these tests locally.
-
由 Heikki Linnakangas 提交于
The test was sensitive to the number of pages in the pg_rewrite system table's index, for no good reason. Also, don't create a new database for it, to speed it up.
-
由 Heikki Linnakangas 提交于
It was only used by the filerep code. Now that that's gone, this was just dead code.
-
由 Jacob Champion 提交于
The DB_IN_STANDBY_NEW_TLI_SET state doesn't really seem to do anything anymore, as of commit 813b817cc. Remove it entirely to get rid of an assertion during standby tests. Also remove multipass function declarations; they're gone.
-