On Mon, 15 Nov 2021 at 12:40, Dilip Kumar wrote:
>
> On Mon, Nov 15, 2021 at 4:46 PM Rafia Sabih wrote:
> >
> > On Mon, 15 Nov 2021 at 10:24, Dilip Kumar wrote:
> > >
> > > On Wed, Nov 10, 2021 at 1:47 PM Rafia Sabih
> > > wrote:
> > > &
On Mon, 15 Nov 2021 at 10:24, Dilip Kumar wrote:
>
> On Wed, Nov 10, 2021 at 1:47 PM Rafia Sabih wrote:
> >
> > > It seems that in beentry->st_idle_time, you want to compute the
> > > STATE_IDLE, but that state is not handled in the outer "if&qu
me which files should I take a look at?
>
> Thanks in advance!
I think you can start by looking into the function grouping_planner
and explore further by going through the route of it.
--
Regards,
Rafia Sabih
On Wed, 10 Nov 2021 at 09:05, Dilip Kumar wrote:
>
> On Tue, Nov 9, 2021 at 8:28 PM Rafia Sabih wrote:
> >
> > On Tue, 2 Nov 2021 at 09:00, Dilip Kumar wrote:
> > >
>
> > > About the patch, IIUC earlier all the idle time was accumulated in the
> &g
On Tue, 2 Nov 2021 at 09:00, Dilip Kumar wrote:
>
> On Tue, Oct 26, 2021 at 5:17 PM Rafia Sabih wrote:
> >
> > >
> > > To provide this information I was digging into how the statistics
> > > collector is working and found out there is already information
On Fri, 22 Oct 2021 at 10:22, Rafia Sabih wrote:
>
> Hello there hackers,
>
> We at Zalando have faced some issues around long running idle
> transactions and were thinking about increasing the visibility of
> pg_stat_* views to capture them easily. What I found
]. Ideally, this would be the values we would
like to see per process in pg_stat_activity.
Curious to know your thoughts on this.
[1]https://github.com/postgres/postgres/blob/cd3f429d9565b2e5caf0980ea7c707e37bc3b317/src/backend/utils/activity/backend_status.c#L593
--
Regards,
Rafia Sabih
On Mon, 14 Jun 2021 at 21:04, Robert Haas wrote:
>
> On Mon, Jun 14, 2021 at 8:51 AM Rafia Sabih wrote:
> > I have a doubt regarding the positioning of clientAuthentication hook
> > in function ClientAuthentication. Particularly, in case of hba errors,
> > i.e. cases uaRe
process the hook function first and then error out...?
--
Regards,
Rafia Sabih
; we better run pg_indent, for all the parallel copy patches once and
> for all, maybe just before we kind of finish up all the code reviews.
>
> > + if (mode == 1)
> > + {
> > + cstate->pcdata->curr_data_block = data_block;
> > + cstate->raw_buf_index = 0;
> > + }
> > + else if(mode == 2)
> > + {
> > Could use macros for 1 & 2 for better readability.
>
> Done.
>
> >
> > +
> > + if (following_block_id == -1)
> > + break;
> > + }
> > +
> > + if (following_block_id != -1)
> > +
> > pg_atomic_add_fetch_u32(_info->data_blocks[following_block_id].unprocessed_line_parts,
> > 1);
> > +
> > + *line_size = *line_size +
> > tuple_end_info_ptr->offset + 1;
> > + }
> > We could calculate the size as we parse and identify one record, if we
> > do that way this can be removed.
> >
>
> Done.
Hi Bharath,
I was looking forward to review this patch-set but unfortunately it is
showing a reject in copy.c, and might need a rebase.
I was applying on master over the commit-
cd22d3cdb9bd9963c694c01a8c0232bbae3ddcfb.
--
Regards,
Rafia Sabih
benefit it gives in
terms of accuracy.
- I would also be interested in understanding if there are cases when
adding this extra step doesn’t help and have you excluded them already
or if some of them are easily identifiable at this stage...?
- is there any limit on the number of columns for which this will
work, or should there be any such limit...?
--
Regards,
Rafia Sabih
it.
-GetPublicationRelations(Oid pubid)
+GetPublicationRelations(Oid pubid, bool include_partitions)
How about having an enum here with INCLUDE_PARTITIONS,
INCLUDE_PARTITIONED_REL, and SKIP_PARTITIONS to address the three
possibilities and avoiding reiterating through the list in
pg_get_publication_tables().
--
Regards,
Rafia Sabih
rel->rd_rel->relreplident == REPLICA_IDENTITY_FULL ||
> - rel->rd_rel->relreplident == REPLICA_IDENTITY_INDEX);
> + rel->rd_rel->relreplident == REPLICA_IDENTITY_INDEX ||
> + rel->rd_rel->relreplident == REPLICA_IDENTITY_NOTHING);
>
> /* use Oid as relation identifier */
> pq_sendint32(out, RelationGetRelid(rel));
>
+1
--
Regards,
Rafia Sabih
or all tables options.
--
Regards,
Rafia Sabih
dy mentioned, there is a rebase required
particularly for 0003 and 0004.
--
Regards,
Rafia Sabih
On Thu, 5 Dec 2019 at 14:45, Robert Haas wrote:
>
> On Tue, Dec 3, 2019 at 8:36 AM Rafia Sabih wrote:
> > While going through this file I noticed some inconsistencies in the
> > comments. Please find attachment for the fix.
>
> Committed. I think only the duplicat
Hello,
While going through this file I noticed some inconsistencies in the
comments. Please find attachment for the fix.
--
Regards,
Rafia Sabih
diff --git a/src/include/executor/instrument.h b/src/include/executor/instrument.h
index 70d8632305..13aaf05453 100644
--- a/src/include/executor
, how can I inform optimizer that
> parallel scan is not possible (because table data is local to the backend)?
>
> One moment, isn't that parallel scans are already restricted for temp
tables, or I have misunderstood something here...?
--
Regards,
Rafia Sabih
, how can I inform optimizer that
> parallel scan is not possible (because table data is local to the backend)?
>
> How about setting parallel_setup_cost to disable_cost in costsize.c for
your specific scan method.
--
Regards,
Rafia Sabih
On Sat, 9 Nov 2019 at 12:52, Andreas Karlsson wrote:
> On 11/8/19 2:52 PM, Rafia Sabih wrote:
> > Now, my question is shouldn't we always use this list in sorted order,
> > in other words can there be scenarios where such a sorting will not
> > help? I am talking about
to accomplish
this.
So, your thoughts, opinions, suggestions are more than welcome.
--
Regards,
Rafia Sabih
node even
without verbose, but for scans it is only with verbose. Am I missing
something or there is something behind? However, I am not sure if this is
the introduced by this patch-set.
--
Regards,
Rafia Sabih
s expecting more like an error at
subscriber saying the table type is not same.
Please find the attached file for the test case, in case something is
unclear.
--
Regards,
Rafia Sabih
-- publisher
create table t (i int, j int, k text) partition by range(i);
create table child1 partition of
ables were to be replicated, but with the extension of the scope of
logical replication to more objects such checks would be helpful.
On a separate note was thinking for partitioned tables, wouldn't it be
cleaner to have something like you create only partition table at the
subscriber and then when logical replication starts it creates the child
tables accordingly. Or would that be too much in future...?
--
Regards,
Rafia Sabih
right database or host and it has nothing to
> do with initialization.
>
> Maybe something like:
>
> "pgbench_accounts is missing, perhaps you need to initialize (\"pgbench
> -i\") in database \"%s\"\n"
>
> The two sentences approach has the logic of "error" and a separate "hint"
> which is often used.
>
+1 for error and hint separation.
--
Regards,
Rafia Sabih
sn't look quite right. I mean to say
> whatever we want to say via this message is correct, but I am not
> completely happy with the display part. How about something like:
> "pgbench_accounts is missing, you need to do initialization (\"pgbench
> -i\") in database \"%s\"\n"? Feel free to propose something else on
> similar lines? If possible, I want to convey this information in a single
> sentence.
>
> How about, "pgbench_accounts is missing, initialize (\"pgbench -i\") in
database \"%s\"\n"?
>
--
Regards,
Rafia Sabih
sults, with a visual
> representation of which GUCs affect which queries.
>
> Wow, that sounds like an elaborate experiment. But where is this
spreadsheet you mentioned ?
--
Regards,
Rafia Sabih
mation is first entered in this structure before actually creating it.
> I think it is extensible to other partitioning schemes as well. Also it
> > is likely to have a positive impact on the queries, because there will
> > be required partitions only and would not require to educate
> > planner/executor about many empty partitions.
>
> Yep, but it creates other problems to solve…
>
> Isn't it always the case. :)
--
Regards,
Rafia Sabih
d it's
interval and start and end values are specified and it creates only the
parent table just like it works today. Now, if there comes a insertion
that does not belong to the existing (or any, in the case of first
insertion) partition(s), then the corresponding partition is created, I
think it is extensible to other partitioning schemes as well. Also it is
likely to have a positive impact on the queries, because there will be
required partitions only and would not require to educate planner/executor
about many empty partitions.
[1]
https://www.postgresql.org/message-id/flat/20190820205005.GA25823%40alvherre.pgsql#c67245b98e2cfc9c3bd261f134d05368
--
Regards,
Rafia Sabih
default is something that remains there by default, isn't
it?
I will be happy to know your thoughts on this.
--
Regards,
Rafia Sabih
things I noticed in the first look,
+/*
+ * When a table has no indexes, save the progress every 8GB so that we can
+ * resume vacuum from the middle of table. When table has indexes we save it
+ * after the second heap pass finished.
+ */
+#define VACUUM_RESUME_BLK_INTERVAL 1024 /* 8MB */
Discrepancy with the memory unit here.
/* No found valid saved block number, resume from the first block */
Can be better framed.
--
Regards,
Rafia Sabih
king BTreeTupleIsPosting to ensure that
+ * it will get what he expects
+ */
Everything reads just fine without 'us'.
/*
+ * This field helps us to find beginning of the remaining tuples from
+ * postings which follow array of offset numbers.
+ */
--
Regards,
Rafia Sabih
d describeTablespaces, precisely
if (verbose && pset.sversion >= 90200)
+ {
appendPQExpBuffer(,
",\n pg_catalog.pg_size_pretty(pg_catalog.pg_tablespace_size(oid))
AS \"%s\"",
gettext_noop("Size"));
+ sizefunc = "pg_catalog.pg_tablespace_size(oid)";
+ }
in describeTablespaces but
if (verbose && pset.sversion >= 80200)
+ {
appendPQExpBuffer(,
",\n CASE WHEN pg_catalog.has_database_privilege(d.datname,
'CONNECT')\n"
"THEN
pg_catalog.pg_size_pretty(pg_catalog.pg_database_size(d.datname))\n"
"ELSE 'No Access'\n"
" END as \"%s\"",
gettext_noop("Size"));
+ sizefunc = "pg_catalog.pg_database_size(d.datname)";
+ }
in listAllDbs.
--
Regards,
Rafia Sabih
On Thu, 25 Jul 2019 at 16:44, Tom Lane wrote:
>
> Rafia Sabih writes:
> > On Thu, 25 Jul 2019 at 13:50, vignesh C wrote:
> >>> Initdb fails when following p
On Thu, 25 Jul 2019 at 13:50, vignesh C wrote:
>
> On Thu, Jul 25, 2019 at 4:52 PM Rafia Sabih wrote:
> >
> > On Thu, 25 Jul 2019 at 07:39, vignesh C wrote:
> > >
> > > Hi,
> > >
> > >
is is already known.
I am also not sure if it is known or intentional. On the other hand I
also don't know if it is practical to give such long names for
database directory anyway.
--
Regards,
Rafia Sabih
wonder about those 'some' reasons.
Caller is responsible for checking BTreeTupleIsPosting to ensure that
+ * he will get what he expects
This can be re-framed to make the caller more gender neutral.
Other than that, I am curious about the plans for its backward compatibility.
--
Regards,
Rafia Sabih
execution
time it starts with the plans one by one, starting from the lower
selectivity one till the query execution completes. It might sound
like too much work but it isn't, there are some theoretical guarantees
to bound the worst case execution time. The trick is in choosing the
plan-set and switching at the time of execution. Another good point
about this is that it works smoothly for join predicates as well.
Since, we are talking about this problem here, I though it might be a
good idea to shed some light on such an approach and see if there is
some interesting trick we might use.
[1] https://dsl.cds.iisc.ac.in/publications/conference/bouquet.pdf
--
Regards,
Rafia Sabih
erly it
> intentionally excludes backwards scanning. I don't see any particular
> reason why that ought to be the case, and it seems like an odd
> limitation for the feature should it be merged. Should that be a
> blocker to merging?
Regarding this, I came across this,
/*
* Incremental so
rows)
Notice that in the explain statement it shows correct number of rows
to be skipped.
--
Regards,
Rafia Sabih
ou want it to work.
--
Regards,
Rafia Sabih
On Mon, 6 May 2019 at 22:39, Ashwin Agrawal wrote:
>
> On Mon, May 6, 2019 at 7:14 AM Andres Freund wrote:
> >
> > Hi,
> >
> > On May 6, 2019 3:40:55 AM PDT, Rafia Sabih
> > wrote:
> > >I was trying the toyam patch and on make check it failed with
On Mon, 6 May 2019 at 16:14, Andres Freund wrote:
>
> Hi,
>
> On May 6, 2019 3:40:55 AM PDT, Rafia Sabih wrote:
> >On Tue, 9 Apr 2019 at 15:17, Heikki Linnakangas
> >wrote:
> >>
> >> On 08/04/2019 20:37, Andres Freund wrote:
> >> >
o be optional.
>
I was trying the toyam patch and on make check it failed with
segmentation fault at
static void
toyam_relation_set_new_filenode(Relation rel,
char persistence,
TransactionId *freezeXid,
MultiXactId *minmulti)
{
*freezeXid = InvalidTransactionId;
Basically, on running create table t (i int, j int) using toytable,
leads to this segmentation fault.
Am I missing something here?
--
Regards,
Rafia Sabih
On Fri, 3 May 2019 at 16:27, Justin Pryzby wrote:
>
> On Fri, May 03, 2019 at 02:55:47PM +0200, Rafia Sabih wrote:
> > On Mon, 22 Apr 2019 at 17:49, Justin Pryzby wrote:
> > >
> > > It's deliberate that \dt doesn't show toast tables.
> > > \d sho
>
> postgres=# \d pg_toast.pg_toast_2600
> TOAST table "pg_toast.pg_toast_2600"
>Column | Type
> +-
> chunk_id | oid
> chunk_seq | integer
> chunk_data | bytea
> Indexes:
> "pg_toast_2600_index" PRIMARY KEY, btree (chunk_id, chunk_seq)
>
+1.
--
Regards,
Rafia Sabih
On Sat, 27 Apr 2019 at 06:18, David Fetter wrote:
>
> On Fri, Apr 26, 2019 at 04:22:18PM +0200, Rafia Sabih wrote:
> > On Fri, 26 Apr 2019 at 14:49, Rafia Sabih wrote:
> > >
> > > On Wed, 24 Apr 2019 at 10:30, Fabien COELHO wrote:
> > > >
> >
tion.
>
> I wonder if there's a different hook I can use when there's a gather node? or
> other ways to get around?
A bit more detail on what are you trying to do exactly like what is
the query you or code you are dealing with, will be helpful in
providing suggestions,
--
Regards,
Rafia Sabih
On Fri, 26 Apr 2019 at 14:49, Rafia Sabih wrote:
>
> On Wed, 24 Apr 2019 at 10:30, Fabien COELHO wrote:
> >
> >
> > Hello David,
> >
> > >>> I noticed that there wasn't a bulk way to see table logged-ness in psql,
> > >>> so I made it
t;
> Maybe you could consider adding a case for prior 9.1 version, something
> like:
>... case c.relistemp then 'temporary' else 'permanent' end as ...
>
>
I was reviewing this patch and found a bug,
create table t (i int);
create index idx on t(i);
\di+
psql: print.c:3452: printQuery: Assertion `opt->translate_columns ==
((void *)0) || opt->n_translate_columns >= cont.ncolumns' failed.
--
Regards,
Rafia Sabih
[3] and [4], cstore_fdw [5]
> were refered to distill down objectives and come up with design and
> implementations to avoid any earlier concerns raised. Learnings from
> Greenplum Database column store also leveraged while designing and
> implementing the same.
>
> Credit: Design is moslty brain child of Heikki, or actually his
> epiphany to be exact. I acted as idea bouncing board and contributed
> enhancements to the same. We both are having lot of fun writing the
> code for this.
>
>
> References
> 1] https://github.com/greenplum-db/postgres/tree/zedstore
> 2]
> https://www.postgresql.org/message-id/flat/CAJrrPGfaC7WC9NK6PTTy6YN-NN%2BhCy8xOLAh2doYhVg5d6HsAA%40mail.gmail.com
> 3]
> https://www.postgresql.org/message-id/flat/20150611230316.GM133018%40postgresql.org
> 4]
> https://www.postgresql.org/message-id/flat/20150831225328.GM2912%40alvherre.pgsql
> 5] https://github.com/citusdata/cstore_fdw
> 6]
> https://www.postgresql.org/message-id/flat/CAOykqKfko-n5YiBJtk-ocVdp%2Bj92Apu5MJBwbGGh4awRY5NCuQ%40mail.gmail.com
> 7]
> https://www.postgresql.org/message-id/d0fc97bd-7ec8-2388-e4a6-0fda86d71a43%40iki.fi
>
>
--
Regards,
Rafia Sabih
invite confusion with "zheap", especially
> in parts of the world where the last letter of the alphabet is
> pronounced "zed," where people are going to say zed-heap and
> zed-store. Brr.
>
+1 on Brr. Looks like Thomas and your thought on having 'z' makes things
popular/stylish, etc. is after all true, I was skeptical back then.
--
Regards,
Rafia Sabih
sign is moslty brain child of Heikki, or actually his
> epiphany to be exact. I acted as idea bouncing board and contributed
> enhancements to the same. We both are having lot of fun writing the
> code for this.
>
>
> References
> 1] https://github.com/greenplum-db/postgres/tree/zedstore
> 2]
> https://www.postgresql.org/message-id/flat/CAJrrPGfaC7WC9NK6PTTy6YN-NN%2BhCy8xOLAh2doYhVg5d6HsAA%40mail.gmail.com
> 3]
> https://www.postgresql.org/message-id/flat/20150611230316.GM133018%40postgresql.org
> 4]
> https://www.postgresql.org/message-id/flat/20150831225328.GM2912%40alvherre.pgsql
> 5] https://github.com/citusdata/cstore_fdw
> 6]
> https://www.postgresql.org/message-id/flat/CAOykqKfko-n5YiBJtk-ocVdp%2Bj92Apu5MJBwbGGh4awRY5NCuQ%40mail.gmail.com
> 7]
> https://www.postgresql.org/message-id/d0fc97bd-7ec8-2388-e4a6-0fda86d71a43%40iki.fi
>
>
Reading about it reminds me of this work -- TAG column storage(
http://www09.sigmod.org/sigmod/record/issues/0703/03.article-graefe.pdf ).
Isn't this storage system inspired from there, with TID as the TAG?
It is not referenced here so made me wonder.
--
Regards,
Rafia Sabih
On Fri, 29 Mar 2019 at 22:07, Tomas Vondra
wrote:
> On Wed, Mar 27, 2019 at 09:06:04AM +0100, Rafia Sabih wrote:
> >On Tue, 26 Mar 2019 at 21:04, Tomas Vondra
> >wrote:
> >
> >> On Mon, Mar 18, 2019 at 11:31:48AM +0100, Rafia Sabih wrote:
> >> >
On Tue, 26 Mar 2019 at 21:04, Tomas Vondra
wrote:
> On Mon, Mar 18, 2019 at 11:31:48AM +0100, Rafia Sabih wrote:
> >On Sun, 24 Feb 2019 at 00:06, Tomas Vondra
> wrote:
> >>
> >> Hi,
> >>
> >> attached is an updated patch, fixing and slightly twea
to solve, I'd not like to add the progress counter for
> >> the sorting tuples.
> >>
> >>
> >>>>- Progress counter for "6. rebuilding index" phase
> >>>> - Should we add "index_vacuum_count" in the view like a v
On Sun, 24 Feb 2019 at 00:06, Tomas Vondra wrote:
>
> Hi,
>
> attached is an updated patch, fixing and slightly tweaking the docs.
>
>
> Barring objections, I'll get this committed later next week.
>
I was having a look at this patch, and this kept me wondering,
+static void
GETARG_INT32(1);
Looks like PG indentation is not followed here for n.
--
Regards,
Rafia Sabih
EnterpriseDB: http://www.enterprisedb.com/
On Mon, Sep 10, 2018 at 4:10 PM, Christoph Berg wrote:
> Re: Rafia Sabih 2018-09-10 8hfpxsrr6mxab54ebju...@mail.gmail.com>
> > Thanks for the interest and help you offered. But I didn't quite get it,
> I
> > tried maintaner-clean and distclean but it didn't work.
>
&g
On Mon, Sep 10, 2018 at 2:56 PM, Daniel Gustafsson wrote:
> > On 10 Sep 2018, at 11:22, Rafia Sabih
> wrote:
> >
> > Hi all,
> >
> > After pulling the the latest commit --
> > e3d77ea6b4e425093db23be492f236896dd7b501,
> I am getting followi
about that...
--
Regards,
Rafia Sabih
EnterpriseDB: http://www.enterprisedb.com/
On Fri, Aug 31, 2018 at 4:11 PM, Maksim Milyutin
wrote:
> On 08/31/2018 01:12 PM, Rafia Sabih wrote:
>
>
>
> On Fri, Aug 31, 2018 at 3:30 PM, Maksim Milyutin
> wrote:
>
>> On 08/31/2018 11:55 AM, Rafia Sabih wrote:
>>
>>
>> Adding to that this p
On Fri, Aug 31, 2018 at 3:30 PM, Maksim Milyutin
wrote:
> On 08/31/2018 11:55 AM, Rafia Sabih wrote:
>
>
> Adding to that this patch needs a rebase. And, please don't forget to run
> 'git diff --check', as it has some white-space issues.
>
>
> git apply /home/ed
get to run
'git diff --check', as it has some white-space issues.
--
Regards,
Rafia Sabih
EnterpriseDB: http://www.enterprisedb.com/
?
>
>
As far as this setting goes, there wasn't any other query using
partition-wise-agg, so, no.
BTW, just an FYI, this experiment is on scale factor 20.
--
Regards,
Rafia Sabih
EnterpriseDB: http://www.enterprisedb.com/
18_pwa_off.out
Description: Binary data
d commit 4c728f382970 forget to make
> sure the null terminator is copied? See attached proposed fix.
>
> Yes, I missed that.
Your patch looks good to me, thanks.
--
Regards,
Rafia Sabih
EnterpriseDB: http://www.enterprisedb.com/
ain analyse output for
either values of parallel_leader_participation and patches.
--
Regards,
Rafia Sabih
EnterpriseDB: http://www.enterprisedb.com/
with parallel_leader_participation = 1;
67 matches
Mail list logo