Re: [HACKERS] [POC] Faster processing at Gather node

2017-11-15 Thread Robert Haas
ny more times),1,1,1,2,2,2,2,2, If we could somehow scramble the distribution of tuples to workers so that this didn't happen, I think it would fix this problem. Neither (2) nor (3) seem terribly easy to implement so maybe we should just go with (1), but I feel like that's not a very deep so

Re: [HACKERS] Parallel Hash take II

2017-11-15 Thread Robert Haas
ead of max_connections. You can't use parallel query for every query at the same time with reasonable settings... -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] Issues with logical replication

2017-11-15 Thread Robert Haas
that, I understand how the lack of that fix could cause: 2017-10-02 18:40:26.101 MSK [2954] ERROR: attempted to lock invisible tuple But I do not understand how it could cause: #3 0x0086ac1d in XactLockTableWait (xid=0, rel=0x0, ctid=0x0, oper=XLTW_None) at lmgr.c:582 -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] Inconsistencies between pg_settings and postgresql.conf

2017-11-15 Thread Robert Haas
sample file intended to be read by humans. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: Further simplification of c.h's #include section

2017-11-15 Thread Robert Haas
uot;win32_more.h"? Seems morally equivalent to what you had before. I think what I would be looking for is a filename that somehow conveys what the difference is between what should go in the existing file and what should go in the new file. If we don't know, maybe we should find out befor

Re: [HACKERS] Re: protocol version negotiation (Re: Libpq PGRES_COPY_BOTH - version compatibility)

2017-11-15 Thread Robert Haas
n. Here's an updated version with that change and a proposed commit message. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company pgwire-be-rmh-v2.patch Description: Binary data

Re: Further simplification of c.h's #include section

2017-11-15 Thread Robert Haas
On Wed, Nov 15, 2017 at 5:22 PM, Tom Lane <t...@sss.pgh.pa.us> wrote: > Thoughts? Sure, having a win32_port.h as a sub-include of port.h seems fine. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] More stats about skipped vacuums

2017-11-27 Thread Robert Haas
On Sun, Nov 26, 2017 at 3:19 PM, Tom Lane <t...@sss.pgh.pa.us> wrote: > Robert Haas <robertmh...@gmail.com> writes: >> On Sat, Nov 25, 2017 at 12:09 PM, Tom Lane <t...@sss.pgh.pa.us> wrote: >>> Mumble. It's a property I'm pretty hesitant to give up, espec

Re: [HACKERS] Challenges preventing us moving to 64 bit transaction id (XID)?

2017-11-27 Thread Robert Haas
, because it would break pg_upgrade. If we get the heap storage API working, then we could have a heap AM that works as it does today and a newheap AM with such changes, but I have a bit of a hard time imagining a patch that causes a hard compatibility break ever being accepted. -- Robert

Re: default range partition and constraint exclusion

2017-11-27 Thread Robert Haas
s working as expected, so that would say we ought to fix the constraint. I am out of time for today but will try to look at this some more tomorrow. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] postgres_fdw bug in 9.6

2017-11-28 Thread Robert Haas
dca96ca4770552a10a6d5af24. We've put an awful lot of effort into this over the last few years; I'm a bit hesitant to believe none of that did anything. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [JDBC] [HACKERS] Channel binding support for SCRAM-SHA-256

2017-11-28 Thread Robert Haas
hink that would be a good thing to have. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: PATCH: pgbench - option to build using ppoll() for larger connection counts

2017-11-28 Thread Robert Haas
r set to a value much less than the number of FDs the system could handle -- but if it is, then this seems like a reasonable fix, once we agree on exactly how the details should look. I don't remember off-hand whether you've done testing to see how poll()/ppoll() performs compares to select(),

Re: ERROR: too many dynamic shared memory segments

2017-11-28 Thread Robert Haas
| error_queue_space != NULL); for (i = 0; i < pcxt->nworkers; ++i) { char *start; With that fix in place, I then hit a crash in parallel bitmap heap scan. After applying no-pstate.patch, which I just committed and back-patched to v10, then things look OK. I'm going to

Re: default range partition and constraint exclusion

2017-11-28 Thread Robert Haas
On Mon, Nov 27, 2017 at 4:01 PM, Robert Haas <robertmh...@gmail.com> wrote: > I am out of time for today but will try to look at this some more tomorrow. Upon closer study this seems to definitely be a correct fix, so I have committed it. Apologies for my earlier confusion. -- Ro

Re: explain analyze output with parallel workers - question about meaning of information for explain.depesz.com

2017-11-28 Thread Robert Haas
eing any unexpected values? FWIW, this looks sensible to me. Not sure if there's any good way to write a regression test for it. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: explain analyze output with parallel workers - question about meaning of information for explain.depesz.com

2017-11-28 Thread Robert Haas
nformation separately when VERBOSE is used has been discussed, and I think that's a good idea, but it may not be enough. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: Fix a typo in xact.c

2017-11-28 Thread Robert Haas
On Mon, Nov 27, 2017 at 8:29 PM, Masahiko Sawada <sawada.m...@gmail.com> wrote: > Attached patch for $subject. > > s/recoreds/record/ Committed. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: ERROR: too many dynamic shared memory segments

2017-11-28 Thread Robert Haas
pretend we didn't? Why not just avoid creating it in the first place, like this? I haven't checked whether this fixes the bug, but if it does, we can avoid introducing an extra branch in BitmapHeapNext. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: pgindent run?

2017-11-28 Thread Robert Haas
On Tue, Nov 28, 2017 at 2:51 PM, Robert Haas <robertmh...@gmail.com> wrote: > There are only four source files where more than a dozen lines will be > touched... ...and of course by "four" I mean "five". -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: Typo in ExecBuildSlotPartitionKeyDescription prologue

2017-11-28 Thread Robert Haas
On Thu, Nov 23, 2017 at 2:26 AM, Rushabh Lathia <rushabh.lat...@gmail.com> wrote: > Here is a patch for fixing the function > ExecBuildSlotPartitionKeyDescription() > prologue. Committed. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

pgindent run?

2017-11-28 Thread Robert Haas
up commiting patches to partition.c to fix bugs and add features, and the fact I've missed that a few times is now coming back to haunt me. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: With commit 4e5fe9ad19, range partition missing handling for the NULL partition key

2017-11-28 Thread Robert Haas
On Thu, Nov 23, 2017 at 5:41 AM, Rushabh Lathia <rushabh.lat...@gmail.com> wrote: >> Good point. Updated patch attached. > > Thanks Amit. > > Patch looks good to me. Committed, except I left out the comment tweak, which seemed irrelevant. -- Robert

Re: [HACKERS] Moving relation extension locks out of heavyweight lock manager

2017-11-28 Thread Robert Haas
boid, reloid> and take the result modulo 1024; lock the RELEXTLOCK at that offset in the array. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] Patch: Add --no-comments to skip COMMENTs with pg_dump

2017-11-28 Thread Robert Haas
ooks OK to me, but I'm a bit concerned about whether we have consensus for it. By my count the vote is 6-3 in favor of proceeding. +1: Robins Tharakan, Stephen Frost, David Fetter, Fabrizio Mello, Michael Paquier, Robert Haas -1: David G. Johnston, Tom Lane, Simon Riggs I guess that's probably suffic

Re: [HACKERS] postgres_fdw super user checks

2017-11-28 Thread Robert Haas
other votes (or changes of opinion). -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] PG10 partitioning - odd behavior/possible bug

2017-11-28 Thread Robert Haas
saction, and now() returns the same value both times because that's how now() works, so the second time the range ends up with the lower and endpoints that are equal. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] Assertion failure when the non-exclusive pg_stop_backup aborted.

2017-11-28 Thread Robert Haas
ight have missed the fact that Michael thinks it's ready to go. Anyone interested in taking a look at this one for commit? -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] [POC] Faster processing at Gather node

2017-11-25 Thread Robert Haas
Context is cheaper. Typically you only want to free manually if the freeing would otherwise not happen soon enough. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] More stats about skipped vacuums

2017-11-25 Thread Robert Haas
ot, because that's bound to be pretty expensive even if you eliminate the I/O cost. Taking a consistent snapshot across all 100,000 tables in the database even if we're only ever going to access 5 of those tables doesn't seem like a good or scalable design. -- Robert Haas EnterpriseDB: http://www.enterpr

Re: default range partition and constraint exclusion

2017-11-24 Thread Robert Haas
= 2. It should be possible to dig down in there step by step and figure out where the wheels are coming off -- have you tried to do that? -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] Restrict concurrent update/delete with UPDATE of partition key

2017-11-24 Thread Robert Haas
automated regression test. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] More stats about skipped vacuums

2017-11-24 Thread Robert Haas
d reluctance to make stats tables that have a row per table any wider at all, IIRC. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: PATCH: pgbench - option to build using ppoll() for larger connection counts

2017-11-29 Thread Robert Haas
good bit of what you say here, I don't agree with that. If the member used by ppoll() (or just poll()) has a different name than the one used for select(), it's much easier to, say, grep for everyplace where the field you care about is used. If you use the same name for different things, that doesn

Re: using index or check in ALTER TABLE SET NOT NULL

2017-11-29 Thread Robert Haas
based on an index scan doesn't necessarily seem like a good idea. We have no guarantee at all that the index scan will be faster than scanning the table would have been, and a single table scan can do multiple verification steps if, for example, multiple columns are set NOT NULL at the same time. --

Re: using index or check in ALTER TABLE SET NOT NULL

2017-11-29 Thread Robert Haas
LL in the column when there isn't, and thus failing spuriously). But it seems like it might be useful if someone can figure out the details of how to make it 100% correct; one index lookup is sure to be a lot quicker than a full table scan. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The

Re: es_query_dsa is broken

2017-11-29 Thread Robert Haas
ry compatibility with EState (and maybe other PlanState nodes) in a released branch. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] path toward faster partition pruning

2017-11-29 Thread Robert Haas
On Wed, Nov 29, 2017 at 3:28 PM, Robert Haas <robertmh...@gmail.com> wrote: > On Wed, Nov 22, 2017 at 3:59 AM, Amit Langote > <langote_amit...@lab.ntt.co.jp> wrote: >> It seems I wrote an Assert in the code to support hash partitioning that >> wasn't based on a va

Re: [HACKERS] INSERT ON CONFLICT and partitioned tables

2017-11-30 Thread Robert Haas
k even when the query has no ON CONFLICT clause, which doesn't seem like a good idea. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] Support to COMMENT ON DATABASE CURRENT_DATABASE

2017-11-30 Thread Robert Haas
ESET >> configuration_parameter >> GRANT ... ON DATABASE CURRENT_DATABASE TO role_specification ... >> REVOKE ... ON DATABASE CURRENT_DATABASE FROM ... > > Moved to next CF with same status, "needs review". Patch no longer applies. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] [PROPOSAL] Temporal query processing with range types

2017-11-30 Thread Robert Haas
creating plan nodes suggests that you are approaching this from the point of view of sticking the logic in near the end of the planning process, which I think is not going to work. Whatever is going to happen here needs to happen at path generation time at the latest, or maybe even earlie

Re: Protect syscache from bloating with negative cache entries

2017-11-29 Thread Robert Haas
ill applies, compiles, and got no > reviews. So moved to next CF. I think we have to mark this as returned with feedback or rejected for the reasons mentioned here: http://postgr.es/m/ca+tgmozjn28uyjrq2k+5idhyxwbder68sctoc2p_nw7h7jb...@mail.gmail.com -- Robert Haas EnterpriseDB: http://www.enterpris

Re: [HACKERS] Proposal: Local indexes for partitioned table

2017-11-30 Thread Robert Haas
ren exist, so trying to use that in > the planner without having a flag could cause invalid plans to be > generated (i.e. ones that would cause nonexistent indexes to be > referenced). Did you do it this way due to locking concerns? -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] Refactoring identifier checks to consistently use strcmp

2017-11-30 Thread Robert Haas
stent would help us avoid subtle bugs, but it seems at least as likely to CREATE subtle bugs, and the status quo is AFAICT harming nobody. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] postgres_fdw bug in 9.6

2017-11-30 Thread Robert Haas
e the way to go... assuming that your contention that this won't break anything is indeed correct. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: IndexTupleDSize macro seems redundant

2017-11-30 Thread Robert Haas
lit. Therefore I propose the attached patch instead. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company remove-indextupledsize-v2.patch Description: Binary data

Re: [HACKERS] More stats about skipped vacuums

2017-11-27 Thread Robert Haas
the corresponding hash) is postponed until a counter > in the section is really requested? The new counters need to be > shown in a separate stats view (maybe named pg_stat_vacuum). Still makes the stats file bigger. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] Transactions involving multiple postgres foreign servers

2017-11-27 Thread Robert Haas
> other feature on reader side, for example, atomic visibility. If we > have atomic visibility feature, we also can solve the above problem. +1 to all of that. ...Robert > > -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: Treating work_mem as a shared resource (Was: Parallel Hash take II)

2017-11-25 Thread Robert Haas
ot seem like scalable > approaches, and so they don't particularly appeal to me as projects. > I'd be happy to be shown to be wrong about this. Yeah, I agree that it's a hard problem. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] <> join selectivity estimate question

2017-11-30 Thread Robert Haas
better but still-funky estimates handy? Like an EXPLAIN plan? -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] create_unique_path and GEQO

2017-11-30 Thread Robert Haas
which contexts end up containing which data structures. All of which, I think, is a long-winded way of saying that I'm going to go commit this patch. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] [PATCH] Call RelationDropStorage() for broader range of object drops.

2017-11-30 Thread Robert Haas
t be careful to document the reasons for the changes we make, even seemingly minor ones, if we don't want developers to be guessing in ten years whether those changes were actually safe and correct. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] Bug in to_timestamp().

2017-11-30 Thread Robert Haas
ession test cases in this patch produce the same results as they would on Oracle? -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] PG10 partitioning - odd behavior/possible bug

2017-11-28 Thread Robert Haas
On Tue, Nov 28, 2017 at 6:42 PM, Joe Conway <m...@joeconway.com> wrote: > Yeah, Tom already pointed that out a while back: > > https://www.postgresql.org/message-id/20986.1504478066%40sss.pgh.pa.us Ah, sorry, hadn't seen that. -- Robert Haas EnterpriseDB: http://www.e

Re: [HACKERS] path toward faster partition pruning

2017-11-29 Thread Robert Haas
utates -> commuted * removed wrong /* empty */ comment * Updated expected output. It surprised me a bit that the tests weren't passing as you had them, but the differences I got - all related to mc3p_default - seemed correct to me -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The

Re: [HACKERS] pow support for pgbench

2017-12-04 Thread Robert Haas
of pow() and a double version of pow(), with the two versions having different names If Raúl is happy with only having an integer version, then I suggest that he adopt #1 and call it good. Otherwise, given that Fabien wants the double version, I suggest we call the integer version pow() and the doubl

Re: [HACKERS] postgres_fdw super user checks

2017-12-04 Thread Robert Haas
ck-branch documentation. Suggestions appreciated. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: explain analyze output with parallel workers - question about meaning of information for explain.depesz.com

2017-12-04 Thread Robert Haas
ntation arrays happen in ExecParallelReinitialize rather than ExecParallelFinish, so that we don't spend time doing it unless the Gather is actually re-executed? -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: Add PGDLLIMPORT lines to some variables

2017-12-05 Thread Robert Haas
; there is no reason to not include 9.3. > > TransactionXmin -- This is needed for the newer heap-matches-index > verification check. Again, I would like this on 9.4+, but 9.3+ works > too. > > Note that somebody asked about running it on Windows recently, and on > one other occ

Re: [HACKERS] Moving relation extension locks out of heavyweight lock manager

2017-12-16 Thread Robert Haas
got x10 faster than current HEAD as I > mentioned before, the performance degradation may not have gotten > decreased than I expected even in case 2. > Since my machine doesn't have enough resources the result of clients = > 50 might not be a valid result. I have to admit that result is surprising to me. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: Package version in PG_VERSION and version()

2017-12-16 Thread Robert Haas
On Fri, Dec 15, 2017 at 10:23 AM, Tom Lane wrote: > Christoph Berg writes: >> Re: Michael Paquier 2017-12-15 >>

Re: pg_(total_)relation_size and partitioned tables

2017-12-16 Thread Robert Haas
t the sum of pg_total_relation_size(regclass) across all relations in the database will equal the size of the database itself. Partitions will be counted a number of times equal to their depth in the partitioning hierarchy. However, I understand that I might get outvoted. -- Robert Haas Enterpris

Re: Protect syscache from bloating with negative cache entries

2017-12-16 Thread Robert Haas
problem in need of solving. Doesn't that just mean we need the cache to be bigger, at least temporarily? -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] pow support for pgbench

2017-12-01 Thread Robert Haas
or this matches? -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] Additional logging for VACUUM and ANALYZE

2017-12-01 Thread Robert Haas
fairly easy to break in subtle ways, so I think more test coverage is good. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] Moving relation extension locks out of heavyweight lock manager

2017-12-01 Thread Robert Haas
n locks */ "If we hold a relation extension lock, release it." +/* Number of partitions the shared relation extension lock tables are divided into */ +#define LOG2_NUM_RELEXTLOCK_PARTITIONS 4 +#define NUM_RELEXTLOCK_PARTITIONS (1 << LOG2_NUM_RELEXTLOCK_PARTITIONS) Dead code. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: Unclear regression test for postgres_fdw

2017-12-01 Thread Robert Haas
On Fri, Dec 1, 2017 at 4:01 AM, Antonin Houska <a...@cybertec.at> wrote: > I see no other problems here. Committed, thanks for the report and review. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: Use get_greatest_modulus() in partition_bounds_equal()

2017-12-01 Thread Robert Haas
st modulus from > a partition bound. partition_bounds_equal() should also do the same > for the sake of consistency and to make sure the in future the code > remains sane if we change the way we store greatest modulus in > PartitionBoundInfoData in future. Makes sense. Committed. -- Rob

Re: [HACKERS] Runtime Partition Pruning

2017-12-01 Thread Robert Haas
; Planning time: 0.588 ms > Execution time: 0.043 ms > (5 rows) I think the EXPLAIN ANALYZE input should show something attached to the Append node so that we can tell that partition pruning is in use. I'm not sure if that is as simple as "Run-Time Partition Pruning: Yes" or if

Re: Would a BGW need shmem_access or database_connection to enumerate databases?

2017-12-01 Thread Robert Haas
On Fri, Dec 1, 2017 at 10:04 AM, Chapman Flack <c...@anastigmatix.net> wrote: > Can I call RegisterDynamicBackgroundWorker when not in the postmaster, > but also not in a "regular backend", but rather another BGW? I believe that doing it from another BGW works fine. -- Ro

Re: [HACKERS] Custom compression methods

2017-12-01 Thread Robert Haas
ms before deciding what I think. :-) It does seem to me that the patch may be aiming at a relatively narrow target in a fairly large problem space, but I don't know whether to label that as short-sightedness or prudent incrementalism. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Ent

Re: Transform for pl/perl

2017-12-01 Thread Robert Haas
Perhaps a committer will look at it at some point. FWIW, although I like the idea, I'm not going to work on the patch. I like Perl, but I neither know its internals nor use plperl. I hope someone else will be interested. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] Partition-wise aggregation/grouping

2017-12-01 Thread Robert Haas
upData > UpdateStmt > +UpperPathExtraData > UpperRelationKind > UpperUniquePath > UserAuth > > Do we commit this file as part of the feature? Andres and I regularly commit such changes; Tom rejects them. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] Transaction control in procedures

2017-12-01 Thread Robert Haas
hould happen in that case, anyway?). -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] Moving relation extension locks out of heavyweight lock manager

2017-12-01 Thread Robert Haas
On Fri, Dec 1, 2017 at 1:28 PM, Robert Haas <robertmh...@gmail.com> wrote: > [ lots of minor comments ] When I took a break from sitting at the computer, I realized that I think this has a more serious problem: won't it permanently leak reference counts if someone hits ^C or an error occ

Re: BUG #14941: Vacuum crashes

2017-12-01 Thread Robert Haas
and_columns [, ...] ] I don't want to add any more options to the older syntax: VACUUM [ FULL ] [ FREEZE ] [ VERBOSE ] [ ANALYZE ] [ table_and_columns [, ...] ] I am slightly confused as to how we got on to this topic since the subject is "Vacuum crashes", but perhaps Lyes was simply spea

Re: [HACKERS] Custom compression methods

2017-12-01 Thread Robert Haas
s use this new compression method for the type tsvector" > (ALTER TYPE tsvector SET COMPRESSION hyperz) or "let's use this new > compression method for the column tc" (ALTER TABLE ALTER COLUMN tc SET > COMPRESSION hyperz). +1 to this, too. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: PATCH: Exclude unlogged tables from base backups

2017-12-14 Thread Robert Haas
k we just assume missing files are an expected failure mode and silently do nothing if asked to remove them. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] parallel.c oblivion of worker-startup failures

2017-12-12 Thread Robert Haas
test case above to throw a proper ERROR when fork_process() returns -1. Please review... -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company parallel-worker-fork-failure.patch Description: Binary data

Re: Using ProcSignal to get memory context stats from a running backend

2017-12-12 Thread Robert Haas
hole concept, just BTW. As I've said before, I grow weary of asking customers to run gdb. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] Custom compression methods

2017-12-14 Thread Robert Haas
hm but has features which are particularly well-suited to, say, JSON or XML data, because it looks for word boundaries to decide on what strings to insert into the compression dictionary. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] Custom compression methods

2017-12-14 Thread Robert Haas
t; ALTER COLUMN SET COMPRESSION awesome PRESERVE pglz; > > Without PRESERVE it will rewrite the whole relation using new > compression. Also the rewrite removes all unlisted compression options > so their compresssion methods could be safely dropped. That all sounds good. >

Re: [HACKERS] Surjective functional indexes

2017-12-14 Thread Robert Haas
partitioning capabilities improve. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] Proposal: Local indexes for partitioned table

2017-12-18 Thread Robert Haas
hout a matching index get one created. > > 3. ALTER INDEX DETACH is not provided. Therefore: once index is valid, >it remains so forever. > > I think this satisfies all concerns. Sounds great to me. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] Custom compression methods

2017-12-18 Thread Robert Haas
any persona, but so far I'm not finding your reasons convincing... -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: es_query_dsa is broken

2017-12-18 Thread Robert Haas
hat it > failed to handle the case where pei is NULL because parallelism was > inhibited. Here's a new version to fix that. Someone might prefer a > coding with "if" rather than the ternary operator, but I didn't have a > strong preference. Looks OK to me. Committed and

Re: autoprewarm is fogetting to register a tranche.

2017-12-18 Thread Robert Haas
ce too much. It catches something that would otherwise be hard to check. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: Protect syscache from bloating with negative cache entries

2017-12-18 Thread Robert Haas
nly those tables really syscache entries; holding onto things used long in the past doesn't save enough to justify the memory used. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: Top-N sorts verses parallelism

2017-12-19 Thread Robert Haas
On Tue, Dec 19, 2017 at 6:54 AM, Thomas Munro <thomas.mu...@enterprisedb.com> wrote: > On Mon, Dec 18, 2017 at 9:29 AM, Robert Haas <robertmh...@gmail.com> wrote: >> I went through the callers to create_sort_path and the only one that >> looks like it can pass a li

Re: Protect syscache from bloating with negative cache entries

2017-12-19 Thread Robert Haas
gt; 1. The first time we are due to expand the hash table, we check >> whether we can forestall that expansion by doing a cleanup; if so, we >> do that instead. > > And if there's any entry in the removed-reloid hash it is > considered while cleanup. As I say, I don't think there's any need for a removed-reloid hash. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: access/parallel.h lacks PGDLLIMPORT

2017-12-19 Thread Robert Haas
I think waiting for a "really good" use case is too high a bar. I think we only need to think that there is a "reasonable" use case. Accordingly, I pushed a commit adding PGDLLIMPORT to both ParallelWorkerNumber and InitializingParallelWorker. -- Robert Haas Enterp

Re: [HACKERS] parallel.c oblivion of worker-startup failures

2017-12-19 Thread Robert Haas
eat these as error conditions, not try to continue the work. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: Add hint about replication slots when nearing wraparound

2017-12-19 Thread Robert Haas
aparound issues). > > I remember a few discussions where a stale replication slot was actually the > culprit in these situations. > > Something like the attached maybe? +1. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: Using ProcSignal to get memory context stats from a running backend

2017-12-19 Thread Robert Haas
elog/ereport this, you need to pass another argument to MemoryContextStats() or add another memory context method. This is pretty much a textbook example of the wrong way to use a global variable, IMHO. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] UPDATE of partition key

2017-12-15 Thread Robert Haas
TableSlot(), and ExecCloseIndices() operations which are currently performed in CopyFrom() and, by separate code, in ExecEndModifyTable(). -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] UPDATE of partition key

2017-12-15 Thread Robert Haas
On Fri, Dec 15, 2017 at 7:58 AM, Robert Haas <robertmh...@gmail.com> wrote: > Reviewing the preparatory patch: I started another review pass over the main patch, so here are some comments about that. This is unfortunately not a complete review, however. - map = ptr->partition_

Re: [HACKERS] replace GrantObjectType with ObjectType

2017-12-15 Thread Robert Haas
switch/case approach, so it's better to align it that way, I think. > It's weird to have to maintain two different styles. Eh, really? What about the two big arrays at the top of objectaddress.c? (This is just a drive-by comment; I haven't looked at the details of this patch.) -- Robert Haas Enterpri

Re: Top-N sorts verses parallelism

2017-12-15 Thread Robert Haas
ght, or am I still confused? -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

Re: [HACKERS] Proposal: Local indexes for partitioned table

2017-12-15 Thread Robert Haas
id's big concern (which is not without foundation). > I have code for proposals 1 and 2. I don't like proposal 2, and David & > Ashutosh don't like (1). Maybe if we all understand (3) we can agree on > using that one? Yes, it would be nice to achieve some sort of consensus and I th

Re: explain analyze output with parallel workers - question about meaning of information for explain.depesz.com

2017-12-13 Thread Robert Haas
which is what we need from that commit. OK, done. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company

  1   2   3   4   5   6   7   8   9   10   >