threads listlengths 1 2.99k |
|---|
[
{
"msg_contents": "I would like to bring up again the topic of statement-level rollback.\nThis was discussed in some depth at [1]. This patch is not based on\nTsunakawa-san's patch submitted in that thread; although I started from\nit, I eventually removed almost everything and replaced with a\ncompletely diff... |
[
{
"msg_contents": "I've been poking around with a feature I've wanted a number of times\r\nin the past, \"EXPLAIN ALTER TABLE\". The idea is that there are a bunch\r\nof optimizations in ALTER TABLE to minimize the amount of work and\r\nlock levels but it's really hard for users to tell whether they've\r\nwritt... |
[
{
"msg_contents": "Greetings,\n\nPerhaps I'm missing something, but in toast_fetch_datum_slice() there's:\n\n\tAssert(!VARATT_EXTERNAL_IS_COMPRESSED(toast_pointer));\n\nFollowed, not long after, by:\n\n if (VARATT_EXTERNAL_IS_COMPRESSED(toast_pointer))\n SET_VARSIZE_COMPRESSED(result, length + VARHDRS... |
[
{
"msg_contents": "While looking into an issue of Pgpool-II, I found an interesting\nbehavior of a PostgreSQL client.\nBelow is a trace from pgproto to reproduce the client's behavior.\n\nIt starts a transacton.\n\nFE=> Parse(stmt=\"S1\", query=\"BEGIN\")\nFE=> Bind(stmt=\"S1\", portal=\"\")\nFE=> Execute(porta... |
[
{
"msg_contents": "In confg.sgml, in the section about max_stack_depth, there's this sentence:\n\n\"The safety margin is needed because the stack depth is not checked in\nevery routine in the server, but only in key potentially-recursive\nroutines such as expression evaluation.\"\n\nSince the change in expressi... |
[
{
"msg_contents": "I don't know if this applies only to pglogical or logical decoding in\ngeneral. This is on a 9.6.10 provider running pglogical 2.2.0. Subscriber\nhas same versions. We had a replication delay situation this morning,\nwhich I think may have been due to a really long transaction but I've yet... |
[
{
"msg_contents": "Hi,\n\nthe possibility to use polymorphic types is a specific interesting\nPostgreSQL feature. The polymorphic types allows to use almost all types,\nbut when some type is selected, then this type is required strictly without\npossibility to use some implicit casting.\n\nSo if I have a fx(any... |
[
{
"msg_contents": "We've had more buildfarm failures due to hard-coded, short timeouts:\n\nhttp://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=chipmunk&dt=2018-10-13%2021%3A06%3A58\n 10s timeout while running pg_recvlogical\nhttp://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=hornet&dt=2018-12-03%2005%3A... |
[
{
"msg_contents": "The docs say:\nhttps://www.postgresql.org/docs/current/runtime-config-query.html\n|min_parallel_table_scan_size Sets the minimum amount of table data that must be scanned in order for a parallel scan to be considered. [...]\n\nI'd like to set parallel_min_table_size=32MB, but it looks like th... |
[
{
"msg_contents": "Commit 96cdeae07 added toast tables to most catalogs. One disadvantage\nis that the toast declarations require hard-coded oids, even though\nonly shared catalogs actually need stable oids. Now that we can assign\noids on the fly, it makes sense to do so for toast tables as well, as\nin the at... |
[
{
"msg_contents": "I've been experimenting with the task proposed in [1] of expanding\nthe text_ops operator family to include type \"name\" as well as\ncross-type text vs. name operators. These operators would need to\noffer collation-aware sorting, since that's exactly the difference\nbetween text_ops and th... |
[
{
"msg_contents": "Hi list,\n\nI was messing around a bit with BRIN code and found some cleanups to be made:\n1. Remove dead code in brin_form_tuple\n2. Add missing Datum conversion macros\n3. Use AttrNumber type in places using 1-based attribute numbers\n\nThough it's not hard to find cases all over source cod... |
[
{
"msg_contents": "If the passed-in type OID isn't one of the very short list that\nTupleDescInitBuiltinEntry supports, it will silently hand back\na broken TupleDesc, rather than throwing an error. How can\nthis possibly be considered good code?\n\n(So far as I can see, none of the extant callers could hit\ns... |
[
{
"msg_contents": "Attached is a patch that reorganizes how and where collations are looked up.\n\nUntil now, a fmgr C function that makes use of collation information\n(for example varstr_cmp(), str_toupper()) gets passed the collation OID,\nlooks up the collation with pg_newlocale_from_collation(), and then d... |
[
{
"msg_contents": "Hackers,\n\nThe .partial mechanism was added in de768844 to help avoid conflicts\nbetween a newly promoted primary and an old primary that might produce\nthe same WAL segment. This works for a single promotion but can become\nproblematic in HA configurations where there may be several promot... |
[
{
"msg_contents": "We got another report today [1] that seems to be due to the problem\nwe've seen before with failed vacuum truncations leaving corrupt state\non-disk [2]. Reflecting on that some more, it seems to me that we're\nnever going to get to a solution that everybody finds acceptable without\nsome ra... |
[
{
"msg_contents": "Hi,\n\nI'm working on the pluggable storage patch and zheap at the moment. The\nlatter the latter has no alignment padding for byval attributes - since\nwe copy them anyway - and I was looking which places need to be adjusted\nfor that. That lead me to notice that get_rel_data_width() curren... |
[
{
"msg_contents": "Hi,\n\nestimate_rel_size() explains:\n\n\t\t\t/*\n\t\t\t * HACK: if the relation has never yet been vacuumed, use a\n\t\t\t * minimum size estimate of 10 pages. The idea here is to avoid\n\t\t\t * assuming a newly-created table is really small, even if it\n\t\t\t * currently is, because that... |
[
{
"msg_contents": "Hello,\n\na customer recently mentioned that they'd like to be able to see when a\n(md5, scram) role had their password last changed. \n\nUse-cases for this would be issueing an initial password and then later\nmaking sure it got changed, or auditing that all passwords get changed\nonce a yea... |
[
{
"msg_contents": "Hello,\n\nThis candidate patch is kind of a bug-solving patch, it applies to one\nunique file : src/backend/parser/parse_utilcmd.c,\non function transformTableLikeClause.\n\nIts aim is to include the \"top-level\" comment of the object (table,\nview, matview...)\nthat is being copied ; alongs... |
[
{
"msg_contents": "Hi,\n\n> On 4/24/17 22:26, Florin Asavoaie wrote:\n> > If there's nobody against this, I can try to do the patch myself,\n> > doesn't look too difficult (I expect it to simply work by\n> > calling SSL_set_tlsext_host_name(SSL_context, PQhost(conn))) somewhere\n> > in initialize_SSL in fe-secu... |
[
{
"msg_contents": "While using printf to do some good old fashioned print debugging, I\nnoticed that it seemed like my printf statements in the executor were\nsometimes not getting printed to the log at all, and sometimes getting\nprinted out *before* the ones I have in the planner.\n\nUpon further investigatio... |
[
{
"msg_contents": "Hi all,\n\nSince 2dedf4d, recovery.conf is dead and all recovery parameters are now\nGUCs. While looking at a patch to switch primary_conninfo and\nprimary_slot_name to be reloadable, Sergei Kornilov had a very good\npoint that the WAL receiver uses a connection string and a physical slot\nn... |
[
{
"msg_contents": "Hi all,\n\nFunction check_outerjoin_delay() is used to detect whether a qual\nreferencing the given relids must be delayed in application due to the\npresence of a lower outer join.\n\nIf the given relids are empty, we should be able to return from this\nfunction via the same fast path as for... |
[
{
"msg_contents": "When we first put in collations support, we basically punted on teaching\nANALYZE, pg_statistic, and the planner selectivity functions about that.\nThey just use DEFAULT_COLLATION_OID independently of the actual collation\nof the data. I tripped over this while investigating making type \"na... |
[
{
"msg_contents": "By chance I noticed that postgres_fdw's postgresGetForeignPlan() assumes\n--- without any checking --- that the outer_plan it's given for a join\nrelation must have a NestLoop, MergeJoin, or HashJoin node at the top.\nThat's been wrong at least since commit 4bbf6edfb (which could cause\ninser... |
[
{
"msg_contents": "The following bug has been logged on the website:\n\nBug reference: 15548\nLogged by: Hugh Ranalli\nEmail address: hugh@whtc.ca\nPostgreSQL version: 11.1\nOperating system: Ubuntu 18.04\nDescription: \n\nApparently Unicode has two ways of accenting a character: as ... |
[
{
"msg_contents": "Hi,\n\nCraig previously worked on $subject, see thread [1]. A bunch of the\nprerequisite features from that and other related threads have been\nintegrated into PG. What's missing is actually allowing logical\ndecoding on a standby. The latest patch from that thread does that [2],\nbut unf... |
[
{
"msg_contents": "Hello,\n\nPer idea of Nikolay Samokhvalov[1] I propose this patch to add the possibility\nto log all statements from a fraction of transactions.\n\nI have several questions:\n * Do we want this feature?\n * How can I add tests? I seems hard to handle tests when a there is duration\nin the o... |
[
{
"msg_contents": "Hi,\n\nAm I missing something or did\n\ncommit 013ebc0a7b7ea9c1b1ab7a3d4dd75ea121ea8ba7\nAuthor: Teodor Sigaev <teodor@sigaev.ru>\nDate: 2015-09-09 18:43:37 +0300\n\n Microvacuum for GIST\n\n Mark index tuple as dead if it's pointed by kill_prior_tuple during\n ordinary (search) sc... |
[
{
"msg_contents": "Hi pgsql-hackers,\n\nI have an extension for postgresql. The extension works across some\ndatabases, and needs to save some data. The data might be modified\ndynamically by the extension, and all the changed result must be saved. I\nhave considered some methods.\n1. Use GUC\nI find no easy wa... |
[
{
"msg_contents": "Currently, tables provide MVCC access semantics as the only option.\n\nA more complete list of desirable semantics in applications are\n\n* MVCC - provide latest consistent view\n* Historical - keep all old row versions so that queries can access data as\nit used to be\n* TTL=Duration - keep ... |
[
{
"msg_contents": "At present we have a timestamp of 'infinity' and infinite ranges, but no\ninfinite interval\n\nSELECT 'infinity'::timestamp;\nworks\n\nSELECT 'infinity'::interval;\nERROR: invalid input syntax for type interval: \"infinity\"\n\nSeems a strange anomaly that we should fix.\n\n-- \nSimon Riggs ... |
[
{
"msg_contents": "This is a common pattern:\n\n PG_TRY();\n {\n ... code that might throw ereport(ERROR) ...\n }\n PG_CATCH();\n {\n cleanup();\n PG_RE_THROW();\n }\n PG_END_TRY();\n cleanup(); /* the same as above */\n\nI've played with a way to express this more ... |
[
{
"msg_contents": "Hackers,\n\nThe alphabetical ordering of pgarch_readyXlog() means that on promotion\n000000010000000100000001.partial will be archived before 00000002.history.\n\nThis appears harmless, but the .history files are what other potential\nprimaries use to decide what timeline they should pick. T... |
[
{
"msg_contents": "I happened to notice today that the initializer macro for dlist_head\nvariables is\n\n#define DLIST_STATIC_INIT(name) {{&(name).head, &(name).head}}\n\nHowever, all the functions that work with dlists are prepared to handle\na dlist_head that starts out as zeroes, so that this could also be\n... |
[
{
"msg_contents": "This is a mostly cleaned-up version of the test patch I posted\npreviously for floating-point output using the Ryu algorithm.\n\nUpstream source: github.com/ulfjack/ryu/ryu at commit 795c8b57a\n\n From the upstream, I've taken only specific files which are\nBoost-licensed, removed code not of... |
[
{
"msg_contents": "Hi,\n\nnbtxlog.c has a fairly large section of code commented out with #ifdef\nUNUSED. With a node justifying that with:\n\n+ * This section of code is thought to be no longer needed, after\n+ * analysis of the calling paths. It is retained to allow the code\n+ * to be reinstated if ... |
[
{
"msg_contents": "Hi,\n\nCurrently nbtree and hash indexes (and soon gist, where it's missing due\nto oversight) compute an xid that is used to resolve recovery conflicts.\nThey do so by visiting all the heap pages\nxl_btree_delete/xl_hash_vacuum_one_page point to item-by-item.\n\nThat's problematic for two pr... |
[
{
"msg_contents": "Hello,\n\nSince libcrypto.so is implicated, Andres asked me off-list if my\nchanges to random number state initialisation might be linked to\nskink's failures beginning 12 or 15 days ago. It appears not, as it\nwas green for several runs after that commit. Looking at the report:\n\n==2802==... |
[
{
"msg_contents": "Hi,\n\nevery now and then I have to investigate an execution plan that is\nstrange in some way and I can't reproduce the same behavior. Usually\nit's simply due to data distribution changing since the problem was\nobserved (say, after a nightly batch load/update).\n\nIn many cases it however ... |
[
{
"msg_contents": "Hi,\n\nThere are some catalog views which do not show the partitioned table and\nits index entry.\nOne of them is \"pg_indexes\" which failed to show the partitioned index.\nAttached the patch which fixes the same.\n\nOther views such as pg_stat*,pg_statio_* has the same problem for\npartitio... |
[
{
"msg_contents": "We recently ran into a funny situation, where autovacuum would not\nremove very old temp tables. The relfrozenxid of those tables was about\nto reach the max freeze age, so monitoring started to complain. It\nturned out that autovacuum saw that the backendId was used by a live\nbackend ... ... |
[
{
"msg_contents": "I have observed that the following pattern is repeating in our data\nmanagement programs:\n\nUPDATE\n event\nSET\n fuid = ${fuid},\n venue_id = ${venueId},\n url = ${url}\nWHERE\n id = ${id} AND\n fuid IS != ${fuid} AND\n venue_id IS != ${venueId} AND\n url IS DISTINCT FROM ${url};\n\... |
[
{
"msg_contents": "In pursuit of places that might not be on board with non-default\ncollations, I tried sticking\n\n Assert(PG_GET_COLLATION() == C_COLLATION_OID);\n\ninto nameeq() and the other name comparison functions, in my build with\ntype name's typcollation changed to \"C\". I'm down to one place in... |
[
{
"msg_contents": "Hello,\n\nI'd like to propose a patch to log bind parameter values not only when \nlogging duration,\nbut also on error (timeout in particular) or in whatever situation the \nstatement normally gets logged.\nThis mostly could be useful when the request originator doesn't log them \neither, so... |
[
{
"msg_contents": "As pointed out by John Naylor [1], it seems during bootstrap mode, we\nare always creating FSM files which are not required. In commit's\nb9d01fe288 and 3908473c80, we have added some code where we allowed\nthe creation of files during mdopen even if they didn't exist during\nthe bootstrap m... |
[
{
"msg_contents": "Awhile back we noticed that a couple of system catalogs had\nacquired indexes on \"text\" columns, which were unsafe because\ntheir sort order was collation-dependent, so that cloning\ntemplate0 with a different database collation could yield\nbroken indexes. We fixed this in commit 0b28ea79... |
[
{
"msg_contents": "Hey Guys,\n\nI was just playing with exploring joins and plans i came across this\n\ncreate table t1(a int);\ncreate table t2(a int);\ninsert into t1 select (x % 10) from generate_series(1, 100000) x;\ninsert into t2 select (x % 100) from generate_series(1, 100000) x;\n\npgtesting=> analyze t... |
[
{
"msg_contents": "Hello,\n\nI have noticed that since ffa4cbd623, a foreign table that pulls data\nfrom a PROGRAM (in this case an unzip call) will fail if there is a \nLIMIT on the SELECT\n(while succeeding without LIMIT). Below is an example.\n(Table size matters, so larger machines than mine may need more t... |
[
{
"msg_contents": "Hello\n\nFound this on Postgres 9.6, but I think it affects back to 9.4.\n\nI've seen a case where reorderbuffer keeps very large amounts of memory\nin use, without spilling to disk, if the main transaction does little or\nno changes and many subtransactions execute changes just below the\nth... |
[
{
"msg_contents": "Hi all,\n\nWhen you look at Postgres' SQL reference documentation for `GRANT`, the\n`ALL TABLES` clause is explained as :\n\n> ALL TABLES also affects views and foreign tables, just like\nthe specific-object GRANT command.\n\nA colleague of mine was asking himself if it included materialized ... |
[
{
"msg_contents": "On 10/15/18, Tom Lane <tgl@sss.pgh.pa.us> wrote:\n> Andres Freund <andres@anarazel.de> writes:\n>> On 2018-10-15 16:36:26 -0400, Tom Lane wrote:\n>>> We could possibly fix these by changing the data structure so that\n>>> what's in a ScanKeywords entry is an offset into some giant string\n>>>... |
[
{
"msg_contents": "While poking at the signal-reporting bug just pointed out by\nErik Rijkers, I couldn't help noticing how many places we have\nthat are doing some equivalent of this ugly dance:\n\n#if defined(HAVE_DECL_SYS_SIGLIST) && HAVE_DECL_SYS_SIGLIST\n {\n char str2[256];\n\n snp... |
[
{
"msg_contents": "While looking over some recent commits, I noticed there are some code\nlines with a double trailing semicolon instead of a single one. The\nattached fixes these.\n\n-- \n David Rowley http://www.2ndQuadrant.com/\n PostgreSQL Development, 24x7 Support, Training & Services",
... |
[
{
"msg_contents": "Hello Hackers,\n\nI'd like to propose DECLARE STATEMENT syntax in ECPG.\n(I took over this work from Ideriha-san.)\nIn this email, I'd like to answer some major questions\nand attach new patches.\n\nDECLARE STATEMENT syntax allows users to declare an identifier\nfor a SQL statement.\nThe prev... |
[
{
"msg_contents": "Hi Amit,\n(CC: -hackers)\n\nI was just going through some of the tests, when I noticed that the\ntests of partition_info.sql have two typos and that the last set of\ntests is imprecise about the expected behavior of the functions.\n\nDo you think that something like the attached is an improve... |
[
{
"msg_contents": "Hi all.\n\nThere is a demand for ECPG to support UNICODE host variables.\nThis topic has also appeared in thread [1].\nI would like to discuss whether to support in postgres.\n\nDo you have any opinion?\n\nThe specifications and usage described below are the same as in [1].\n\nRequirements\n=... |
[
{
"msg_contents": "Back when Pg added statement-level triggers, I was interested in the\npotential promise of moving referential integrity checks to statement-level\ntriggers.\n\nThe initial conversation, along with Kevin Grittner's POC script (in SQL)\nthat showed a potential for a 98% reduction in time spent ... |
[
{
"msg_contents": "In digging around the codebase (see thread: Referential Integrity Checks\nwith Statement-level Triggers), I noticed that unique constraints are\nsimilarly enforced with a per-row trigger.\n\nThe situation with unique indexes is fairly similar to the situation with\nRI checks: there is some ov... |
[
{
"msg_contents": "Folks,\n\nPlease find attached a run of a tool that looks for duplicated tokens.\nI've removed some things that seem like false positives, basically all\nfrom the stemmer part of the source, but there's still a lot.\n\n- Is it worth trying to track these down and factor them out?\n- Would it ... |
[
{
"msg_contents": "Our deployment script failed to notice dozens of commands failed in a\ntransaction block and I only noticed due to keeping full logs and monitoring\nfor error_severity>'LOG'. I would have thought that exit status would be\nnonzero had an error occured in an earlier script.\n\nThe docs since ... |
[
{
"msg_contents": "Hi all,\n\nFor very large databases, the dbsize function `pg_database_size_name()`\netc. could be quite slow, since it needs to call `stat()` system call on\nevery file in the target database.\n\nWe proposed a new solution to accelerate these dbsize functions which check\nthe disk usage of da... |
[
{
"msg_contents": "Hello.\n\nWe can create a database named \"replication\".\n\n$ createdb replication\n\nA pg_hba.conf entry with DATABASE=\"all\" is described as 'does not\nmatch \"replication\"' in the comment there, but actually it\nmatches and we can connect to the database\n\"replication\". (Documentation... |
[
{
"msg_contents": "I recently stumbled upon the following code in ginfast.c:\n\nwhile (collector->ntuples + nentries > collector->lentuples)\n{\n collector->lentuples *= 2;\n collector->tuples = (IndexTuple *) repalloc(collector->tuples,\n sizeof(IndexTuple) * collector->lentuples);\n}\n\nit does not... |
[
{
"msg_contents": "Hi,\n\nI propose a simple patch (works-for-me), which adds --force (-f)\noption to dropdb utility.\n\nPros: This seems to be a desired option for many sysadmins, as this\nthread proves: https://dba.stackexchange.com/questions/11893/force-drop-db-while-others-may-be-connected\nCons: another p... |
[
{
"msg_contents": "Hi friends,\n\nFirst, I've found cfbot really useful as tool for improving code\ncorrectness. So thanks for that.\n\nSecond, since the project does have a defined style checker, do you\nthink it would be possible to run it as part of cfbot and report errors?\n\nThanks,\n--Robbie",
"msg_d... |
[
{
"msg_contents": "In readfuncs.c, we have READ_ATTRNUMBER_ARRAY, READ_OID_ARRAY,\nREAD_INT_ARRAY, READ_BOOL_ARRAY, but the writing side in outfuncs.c is\ncoded by hand in each case. Any reason for this?\n\nHere is a patch that adds WRITE_ATTRNUMBER_ARRAY, WRITE_OID_ARRAY,\nWRITE_INT_ARRAY, WRITE_BOOL_ARRAY. ... |
[
{
"msg_contents": "With various patches and discussions around collations going on, I\nfigured I'd send in my in-progress patch for insensitive collations.\n\nThis adds a flag \"insensitive\" to collations. Such a collation disables\nvarious optimizations that assume that strings are equal only if they\nare by... |
[
{
"msg_contents": "In another patch discussion it was brought up why the patch doesn't use\nthe IndexIsValid() etc. macros.\n\nThey are defined thus:\n\n/*\n * Use of these macros is recommended over direct examination of the state\n * flag columns where possible; this allows source code compatibility with\n * ... |
[
{
"msg_contents": "While translating the manual into Japanese, I had a hard time to\nparse following sentence in func.sgml:\n\n Note that granting users the EXECUTE privilege on the\n <function>pg_read_file()</function>, or related, functions allows them the\n ability to read any file on the server whi... |
[
{
"msg_contents": "Hi,\n\nRight now there's no easy way to use the compiler to ensure that all\nplaces that need to deal with all kinds of relkinds check a new\nrelkind. I think we should make that easier by moving RELKIND_* to an\nenum, with the existing letters as the value.\n\nObviously we cannot really do ... |
[
{
"msg_contents": "Hi,\n\nThe current pluggable table storage patchset [1] introduces the ability\nto specify the access method of a table (CREATE TABLE ... USING\n\"ident\"). The patchset currently names the current storage method\n(i.e. heapam.c et al) \"heap\" (whereas e.g. zheap's in named, drumroll,\nzheap... |
[
{
"msg_contents": "Hi\n\nsome customer has PostgreSQL 9.5.5. He cannot to login to database due\nbroken index\n\nWhen I use single mode against to this database I got a error\n\nstart transaction\ncould not open critical system index 2679\n\nUnfortunately option -E doesn't show anything\n\nRegards\n\nPavel\n\nH... |
[
{
"msg_contents": "Hi,\n\nThe zheap patchset, even after being based on pluggable storage,\ncurrently has the following condition in RelationAddExtraBlocks():\n\t\tif (RelationStorageIsZHeap(relation))\n\t\t{\n\t\t\tAssert(BufferGetBlockNumber(buffer) != ZHEAP_METAPAGE);\n\t\t\tZheapInitPage(page, BufferGetPage... |
[
{
"msg_contents": "Do you know about GraalVM (https://www.graalvm.org/)? This is a new\npolyglot VM that can run in context of Oracle and MySql, I think that\nsupporting it on Postgresql will be a good thing.\n\nDo you know about GraalVM (https://www.graalvm.org/)? This is a new polyglot VM that can run in cont... |
[
{
"msg_contents": "Hi, hackers. Can you tell me which tools do you prefer when you update *.po translations? Thanks!\n",
"msg_date": "Wed, 19 Dec 2018 18:54:28 +0300",
"msg_from": "=?utf-8?B?0JTQvNC40YLRgNC40Lkg0JLQvtGA0L7QvdC40L0=?=\n <carriingfate92@yandex.ru>",
"msg_from_op": true,
"msg_subje... |
[
{
"msg_contents": "While working on the problem of lowering lock levels for ATTACH\nPARTITION and DETACH PARTITION, I discovered that DETACH PARTITION\ntakes only AccessShareLock on the table being detached, which I think\nis not good. It seems to me that at a minimum it needs to take a\nself-exclusive lock, b... |
[
{
"msg_contents": "Contrary to popular understanding, in classified environments it is common\nto have data marked with a variety of combinations that make it difficult\nto create roles and labels that match the various permutations. As a simple\nexample, a document could be classified as follows:\n\nReleasable... |
[
{
"msg_contents": "We have master server on one Datacenter, and the Replicate is in another\nDatacenter\n\nIntermittently, on the replicate server, we see this as below.\n\n“2018-12-19 07:24:25 UTC: : @: [25775]: [2-1] FATAL: could not receive data\nfrom WAL stream: server closed the connection unexpectedly\nT... |
[
{
"msg_contents": "Hi hackers,\r\n\r\nI've attached a few patches to add some options to vacuumdb that might\r\nbe useful. Specifically, these patches add the --skip-locked,\r\n--min-xid-age, --min-mxid-age, and --min-relation-size options.\r\n\r\nIf an option is specified for a server version that is not supp... |
[
{
"msg_contents": "Greetings,\n\nJust would like to follow up this issue and fix proposal. We really would like to have this issue fixed in PG. Could someone give some suggestions to the fix proposal? Or other ideas to fix this issue?\n\nLooking forward for your feedbacks!\n\n\nBest regards,\n\n--\n\nChengchao ... |
[
{
"msg_contents": "I think this is a bug:\n\npostgres=# alter table only x2 add if not exists i integer;\nERROR: 42701: column \"i\" of relation \"x2\" already exists\n\nNote that it does not occur without the ONLY:\n\npostgres=# alter table x2 add if not exists i integer;\nNOTICE: 42701: column \"i\" of rel... |
[
{
"msg_contents": "As many of you will know, part of my planned work for PG12 is to\nfurther improve the performance of querying partitioned tables. Amit\nLangote is doing quite a bit of work on the planner side of things to\nremove the per-partition overhead to reduce that to just for\npartitions that survive ... |
[
{
"msg_contents": "Back in 2016 [1] there was some discussion about using the POPCNT\ninstruction to improve the performance of counting the number of bits\nset in a word. Improving this helps various cases, such as\nbms_num_members and also things like counting the allvisible and\nfrozen pages in the visibili... |
[
{
"msg_contents": "Hi,\n\nI found the problem about selectivity estimate for range queries\non master as follows. This is my test case:\n\npostgres=# CREATE TABLE test(id int, val text);\nCREATE TABLE\npostgres=# INSERT INTO test SELECT i, 'testval' FROM generate_series(0,29999)i;\nINSERT 0 30000\npostgres=# V... |
[
{
"msg_contents": "de.po's error message when you try to \"REINDEX DATABASE otherdb\" has\nbeen the wrong way round since 2011:\n\ndiff --git a/src/backend/po/de.po b/src/backend/po/de.po\nindex ca52df6731..6aa4354d82 100644\n--- a/src/backend/po/de.po\n+++ b/src/backend/po/de.po\n@@ -7540,7 +7540,7 @@ msgstr \... |
[
{
"msg_contents": "Hi,\n\nCurrently we can skip header line on COPY FROM but having the ability to\nskip and stop copying at any line can use to divide long copy operation and\nenable to copy a subset of the file and skipping footer. Attach is a patch\nfor it\n\n\nRegards\n\nSurafel",
"msg_date": "Thu, 20 D... |
[
{
"msg_contents": "In connection with David Rowley's proposal to change bitmapset.c to use\n64-bit words, I dug out an old test case I had for a complex-to-plan query\n(attached). Andres Freund posted this to the lists perhaps ten years ago,\nthough I can't locate the original posting right now.\n\nI was distr... |
[
{
"msg_contents": "Monitoring progress of CREATE INDEX [CONCURRENTLY] is sure to be welcome,\nso here's a proposal.\n\nThere are three distinct interesting cases. One is straight CREATE\nINDEX of a standalone table; then we have CREATE INDEX CONCURRENTLY;\nfinally, CREATE INDEX on a partitioned table. Note th... |
[
{
"msg_contents": "Check for conflicting queries during replay of gistvacuumpage()\n\n013ebc0a7b implements so-called GiST microvacuum. That is gistgettuple() marks\nindex tuples as dead when kill_prior_tuple is set. Later, when new tuple\ninsertion claims page space, those dead index tuples are physically de... |
[
{
"msg_contents": "Hi all,\n\nAlvaro has cleaned up a couple of error messages recently so as they do\nnot include the function name in what gets translated as per 68f6f2b7.\n\nWhile looking in the code for similar patterns, I have been reminded\nthat pg_stop_backup() is included in some messages when waiting f... |
[
{
"msg_contents": "Hi,\nI want to speed up the creation of UPDATE/DELETE generic plan for tables partitioned into a lot.\n\nCurrently, creating a generic plan of UPDATE/DELTE for such table, planner creates a plan to scan all partitions.\nSo it takes a very long time.\nI tried with a table partitioned into 8192... |
[
{
"msg_contents": "Hi,\n\nI compared INSERT/UPDATE/DELETE/SELECT throughput with PostgreSQL and another dbms.\nFor INSERT/DELETE/UPDATE, PostgreSQL performance is superior, but for SELECT, PostgreSQL performance is slightly lower than another dbms.\n\nBecause information may be missing, it may be difficult, but... |
[
{
"msg_contents": "Is there ever a reason why you would *not* want\nrecovery_target_timeline=latest in standby mode?\n\npg_basebackup -R doesn't set it. That seems suboptimal.\n\nPerhaps this could be the default in standby mode, or even the implicit\ndefault, ignoring the recovery_target_timeline setting alto... |
[
{
"msg_contents": "Hi Hackers,\n\nHere is a patch that passes the -j option from pg_upgrade down to \nvacuumdb if supported.\n\nI'll add it to the January 'Fest.\n\nThanks for considering !\n\nBest regards,\n Jesper",
"msg_date": "Fri, 21 Dec 2018 05:12:16 -0500",
"msg_from": "Jesper Pedersen <jesper.p... |
[
{
"msg_contents": " Hi,\n\nOne consequence of using the \"C\" collation in the catalog versus\nthe db collation is that pg_dump -t with a regexp may not find\nthe same tables as before. It happens when these conditions are\nall met:\n- the collation of the database is not \"C\"\n- the regexp has locale-dependan... |
[
{
"msg_contents": "Hi All,\n\nI'm seeing some very slow queries and it looks like the query planner is\ndeciding to do a 'Nested Loop Left Join'\nwhich is slow. When I SET enable_nestloop=OFF for testing it does a 'Hash\nLeft Join' which is much faster.\n\nI think I can see the cause of the problem in the exa... |
[
{
"msg_contents": "Folks,\n\nI'm trying to compile master (c952eae52a33069e2e92d34f217b43d0eca3d7de)\non Termux, using the supplied settings, as follows.\n\npg_config --configure | xargs ./configure > configure.out 2>configure.err\nmake -j 4 > make.out 2> make.err\n\nThere appears to be some confusion somewhere... |
[
{
"msg_contents": "Hi,\n\nthe attached patch adds offline enabling/disabling of checksums to\npg_verify_checksums. It is based on independent work both Michael\n(Paquier) and me did earlier this year and takes changes from both, see\nhttps://github.com/credativ/pg_checksums and\nhttps://github.com/michaelpq/pg_... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.