threads listlengths 1 2.99k |
|---|
[
{
"msg_contents": "Following belows thread\n\nhttp://www.mail-archive.com/pgsql-hackers@postgresql.org/msg43381.html\n\nI am experiencing the same problem on a brand new, extremely fresh, solaris 5.9 and PostgreSQL 7.4.6 box.\n\nAs expected, the problem was solved when passing the -d [1-5] to the postmaster. Bu... |
[
{
"msg_contents": "> This \"portable\" function is so unportable that I see no reason to\n> accept it as precedent.\n\nHm. Instead of altering the syntax, what slipping in the last\ninserted/updated tuple into the PQResult object? Maybe is a protocol\nlevel option? Now everybody gets to use it with minimal mu... |
[
{
"msg_contents": "> > According to our RELEASE_CHANGES documentation:\n> \t\n> > \tThe major version number should be updated whenever the \n> source of the\n> > \tlibrary changes to make it binary incompatible. Such \n> changes include,\n> > \tbut are not limited to:\n> \t\n> > \t1. Removing a public function... |
[
{
"msg_contents": "Hi list,\n\n\nAttached for your perusal, unicode versions of upper/lower, that work\nindependent of locale except for the following languages:\n\nTurkish, Azeri, and Lithuanian.\nThere are 15 locale specific cases in total not covered.\n\n\n-- \nJohn Hansen <john@geeknet.com.au>\nGeekNET",
... |
[
{
"msg_contents": "This behavior seems inconsistent:\n\nneilc=# create table abc (a int, b int);\nCREATE TABLE\nneilc=# create function foo_abc() returns setof abc as 'select * from\nabc' language sql;\nCREATE FUNCTION\nneilc=# insert into abc values (5, 10);\nINSERT 17234 1\nneilc=# insert into abc values (10,... |
[
{
"msg_contents": "\nA few thoughts on LWLock data structures...\n\nIn lwlock.c we hold a list of lwlocks held:\nheld_lwlocks[MAX_SIMUL_LWLOCKS]\nwhere\n#define MAX_SIMUL_LWLOCKS\t100\n\nThe code for LWLockRelease assumes that the last acquired lock will\nalways be the first one to be released, and uses an O(N)... |
[
{
"msg_contents": "\nFollowing some advice from Intel,\nhttp://www.intel.com/cd/ids/developer/asmo-na/eng/technologies/threading\n/20469.htm?page=2\nI'm looking at whether the LWLock data structures may be within the same\ncache line.\n\nIntel uses 128 byte cache lines on its high end processors.\n\nslru.c uses... |
[
{
"msg_contents": "Hi,\n I am pavan working in database systems lab in Indian Institute of\nscience. As my M.E project I am trying to add new join type to postgresql.\n\n I am searching on net for documentation for developers which will\ndescribe the software architecture of postgres and tell what... |
[
{
"msg_contents": "I've been doing a bit of research on $subj, and coming to the conclusion\nthat the ARC patent is a lot narrower than it might appear. In fact\nmost of the parts of the algorithm that we actually want have prior art.\nI looked in particular at Johnson and Shasha's well-known \"2Q\" paper,\npu... |
[
{
"msg_contents": "I just finally got around to playing with PG 8.0.1.\n\nRegression passes just fine on UnixWare 7.1.4mp2.\n\nWe can *NOT* enable thread safety for lack of a 2 arg version\nof sigwait().\n\nI've asked when it's coming :)\n\nLER\n\n\n-- \nLarry Rosenman http://www.lerctr.org/... |
[
{
"msg_contents": "Marc and/or Bruce:\n\nHey, for my information (people ask me this a lot) can one of you do a count \nof patch submitters for 8.0? For 7.4, it was around 180.\n\n-- \nJosh Berkus\nAglio Database Solutions\nSan Francisco\n",
"msg_date": "Fri, 4 Feb 2005 10:12:27 -0800",
"msg_from": "Jo... |
[
{
"msg_contents": "Michael Klatt reported here:\nhttp://archives.postgresql.org/pgsql-admin/2005-02/msg00031.php\nthat we have problems because the flat files global/pg_pwd\nand global/pg_group aren't rebuilt following WAL recovery.\nThis has in fact been a bug since we created WAL, although it's\ncertainly far... |
[
{
"msg_contents": "Does anyone know what to do with this?\n\nC:\\Downloads\\PostGreSQL\\v8.0.0\\src\\interfaces\\libpq>\"c:\\Program \nFiles\\Borland\\CB\nuilder6\\Bin\\make.exe\" -DCFG=Release /f bcc32.mak\nMAKE Version 5.2 Copyright (c) 1987, 2000 Borland\nBuilding the Win32 DLL and Static Library...\n\nConf... |
[
{
"msg_contents": "Here's one:\nI have the USA census TIGER database loaded, the WHOLE THING, the whole\ncountry. It isn't the biggest database, but it is about 40G before\nindexes. Every table is over a few million rows. I can quite honestly say,\na sequential scan is almost never the right thing to do.\n\nInf... |
[
{
"msg_contents": "Sorry, hit the wrong key.\n\n----- Forwarded message from Alvaro Herrera <alvherre@dcc.uchile.cl> -----\n\nDate: Fri, 4 Feb 2005 22:39:11 -0300\nFrom: Alvaro Herrera <alvherre@dcc.uchile.cl>\nTo: Tom Lane <tgl@sss.pgh.pa.us>\nSubject: Re: [HACKERS] Fixing flat user/group files at database sta... |
[
{
"msg_contents": "> \n> Just an idle thought, but each connection to the DB could add a fixed \n> amount to some queueing parameter. The amount added to be set \n> per backend, \n> and the client could use a SET variable to adjust the \n> standard amount for \n> it's own backend. When the client dies/disconnec... |
[
{
"msg_contents": "\n\nJust brain storming a bit here. It seems to me there are two approaches for\ncross-column statistics within a single table:\n\nA) Treat an index on <a,b,c> the same way Postgres treats an expression index\n on ROW(a,b,c). That is, gather full statistics on the distribution of that\n n... |
[
{
"msg_contents": "I get the following:\n\n$ TRAP: FailedAssertion(\"!(!(tup->t_data->t_infomask & 0x0010))\", File: \"heapam.c\", Line: 1133)\n\n\nwhen I try to cluster this table:\n\nCREATE TABLE virtusers (\n lhs text,\n rhs text,\n insert_date timestamp(0) with time zone DEFAULT now(),\n ins... |
[
{
"msg_contents": "Can anyone think of a reason we aren't inlining MemoryContextSwitchTo()\nin GCC builds, similarly to the way list_head() et al are handled?\n\nIt wouldn't be a huge gain, but I consistently see MemoryContextSwitchTo\neating a percent or three of most profiles.\n\n\t\t\tregards, tom lane\n",
... |
[
{
"msg_contents": "We've been speculating for awhile about ways to reduce contention for\nthe BufMgrLock, in hopes of fixing the \"context swap storm\" problem\nexhibited by certain patterns of concurrent queries. The traditional\nway to fix such problems is to divide what is protected by the lock\ninto finer-... |
[
{
"msg_contents": "Hi\n\nAs VACUUM is not something that can be rolled back, could we not make it\nrun completely outside transactions. It already needs to be run outside\na transaction block.\n\nI try to explain the problem more thoroughly below (I'm quite sleepy, so\nthe explanation may be not too clear ;)\n\... |
[
{
"msg_contents": "For first information, I already have Large Object that work fine on\nPostgreSQL Windows Version. Now I plan to migrate all of Windows version\nto Linux version. But I got some problem that make it cannot work fine\nlike on Windows version.\n \nI found that at least Large Object that now fail... |
[
{
"msg_contents": "Dear Michael Fuhr,\n The word you said is correct on most information. I will give you\nmore information.\n\n<Q> I'd guess that you haven't installed some third-party modules that\nyou need on the Linux box, or that you've installed them in the wrong\nplace.\n<A> I don't make sure about t... |
[
{
"msg_contents": "\nThe following bug has been logged online:\n\nBug reference: 1467\nLogged by: Florian Hars\nEmail address: hars@bik-gmbh.de\nPostgreSQL version: 8.0.1\nOperating system: All\nDescription: fe_connect doesn't handle EINTR right\nDetails: \n\nThe file pgsql/src/inter... |
[
{
"msg_contents": "Hello,\n\nHas anyone seen the following:\n\nhttp://pecl.php.net/package/PDO\n\nThe description from the site:\n\nPDO provides a uniform data access interface, sporting advanced features \nsuch as prepared statements and bound parameters. PDO drivers are \ndynamically loadable and may be devel... |
[
{
"msg_contents": "Hello,\n\nHas anyone seen the following:\n\nhttp://pecl.php.net/package/PDO\n\nThe description from the site:\n\nPDO provides a uniform data access interface, sporting advanced features\nsuch as prepared statements and bound parameters. PDO drivers are\ndynamically loadable and may be develop... |
[
{
"msg_contents": "I was looking at some stubborn queries in one of our applications, and \nnot knowing the internals of the query planner, thought I might ask if \nthis planner improvement is possible at all. We have queries with the \ngeneral form of IN (SELECT FROM AA JOIN (SELECT foo UNION ALL SELECT \nbar)... |
[
{
"msg_contents": "On Mon, Feb 07, 2005 at 02:33:24PM +0700, Premsun Choltanwanich wrote:\n\n> > I'd guess that you haven't installed some third-party modules that\n> > you need on the Linux box, or that you've installed them in the wrong\n> > place.\n>\n> I don't make sure about third-party information cause a... |
[
{
"msg_contents": "Hi everyone,\n\nI had a crash last night and since while vacuuming databases (either full\nor lazy) I get this error:\n\nduplcate key violates unique cnstraint \"pg_statistic_relid_att_index\"\n\nthis is 7.4.6 on unixware.\n\nHow bad is this?\n\nTIA\n\n-- \nOlivier PRENANT \t Te... |
[
{
"msg_contents": "Thank You Michael Fuhr ,\n \nNow, I already installed the contrib/lo and contrib/dbsize modules from\nsource as your suggestion.\n \n>From my first information, I got 3 errors /lo, /dbsize and /admin. That\nmean now only /admin that I cannot found anywhere. \n \nI still try to find it. If an... |
[
{
"msg_contents": "Hi,\n\nJust a quick check that the extension to AT TIME ZONE to allow \nspecifying intervals as well as country/city is on the list for 8.1.\n\nI believe it was a fairly simple thing to do now that we have our own \ntime zone library...\n\nChris\n\n\n",
"msg_date": "Tue, 08 Feb 2005 09:46... |
[
{
"msg_contents": "> Hi,\n> \n> Just a quick check that the extension to AT TIME ZONE to \n> allow specifying intervals as well as country/city is on the \n> list for 8.1.\n> \n> I believe it was a fairly simple thing to do now that we have \n> our own time zone library...\n\nYeah, this is on my personal \"hope... |
[
{
"msg_contents": "A couple of us using the US Census TIGER database have noticed something\nabout the statistics gathering of analyze. If you follow the thread \"Query\nOptimizer 8.0.1\" you'll see the progression of the debate.\n\nTo summarize what I think we've seen:\n\nThe current implementation of analyze ... |
[
{
"msg_contents": "\nShort summary:\n\n * It looks to me like the planner vastly overestimates \n the # of pages read by index scan in quite a few of my\n tables even though stats collected by ANALYZE are correct.\n\n * The problem happens any time you have multiple columns\n that have a number of re... |
[
{
"msg_contents": "\nI believe that this is what Oleg et al tap into with the tsearch2 stuff, \nno? I have someone asking me about it, and want to make sure that I'm \ntelling him the right answer ... is this what GiST is? And, if so, what \nis a *good* doc for me to point them at to get up to speed with it? ... |
[
{
"msg_contents": "A question to the hackers:\n\nIs there a way, and if I'm being stupid please tell me, to use something\nlike a row ID to reference a row in a PostgreSQL database? Allowing the\ndatabase to find a specific row without using an index?\n\nI mean, an index has to return something like a row ID fo... |
[
{
"msg_contents": "Hi,\nA TRUSTED language specifies that ordinary users can use the language. \nIt also implies that access to the file system should be prevented. In \nessence, ordinary users can never access the filesystem.\n\nIs it OK to design a trusted language so that it allows access to the \nfilesystem... |
[
{
"msg_contents": "Mark, Stephen, etc:\n\n> > I can see your point, however I wonder if the issue is that the default\n> > stats settings of '10' (3000 rows, 10 histogram buckets) is too low, and\n> > maybe we should consider making a higher value (say '100') the default.\n>\n> Personally, I think that'd be rea... |
[
{
"msg_contents": "I don't sure. However It maybe possibled cause I use PostgreSQL 8 RC1\nwith pgAdmin.\n\n>>> Michael Fuhr <mike@fuhr.org> 09-Feb-05 11:22 AM >>>\n\nOn Tue, Feb 08, 2005 at 02:55:01PM +0700, Premsun Choltanwanich wrote:\n\n> From my first information, I got 3 errors /lo, /dbsize and /admin.\nTh... |
[
{
"msg_contents": "\n\n\n-----Original Message-----\nFrom: pgsql-hackers-owner@postgresql.org on behalf of Michael Fuhr\nSent: Wed 2/9/2005 4:22 AM\nTo: Premsun Choltanwanich\nCc: pgsql-hackers@postgresql.org\nSubject: Re: [HACKERS] How can I use large object on PostgreSQL Linux\n \n> I wonder if admin is pgadm... |
[
{
"msg_contents": "Those instructions are good but i want to call this function only for \" select consulta_contacto(1)\" nothing more, Is really necesary to use \"AS ...\"?\n\nthis is the call:\n\n\nselect * from consulta_contacto(1) as (cont_codigo integer, \n ... |
[
{
"msg_contents": "I am doing an experiment in which I need the following:\n\nSET enable_mergejoin = false;\nSET enable_hashjoin = false;\n\nSELECT ...\nFROM tab00 as T00, tab01 as T01, tab02 as T02, tab03 as T03 \nWHERE T00.id = T01.id \nAND T00.id = T02.id \nAND T00.id = T03.id \nLIMIT 51;\n\nThere's an index... |
[
{
"msg_contents": "Hello\n\nI have very slow SP this type:\nBEGIN\n CREATE TEMP TABLE xxx ON COMMIT DROP();\n WHILE n > 0 LOOP -- n >> 0\n FOR _r IN EXECUTE 'SELECT ...' LOOP\n \n RETURN NEXT _r;\n EXECUTE 'UPDATE xxx SET item = 1 WHERE id = '||_r.id;\n n := n - 1;\n END LOOP;\n END L... |
[
{
"msg_contents": "How to write a rule to avoid queries like update,insert,delete,select from any database users except superuser. \n \nthanks,\nPrashant\n\n\t\t\n---------------------------------\nDo you Yahoo!?\n Yahoo! Search presents - Jib Jab's 'Second Term'\nHow to write a rule to avoid queries like upda... |
[
{
"msg_contents": "For about 5 years now, I have been using a text search engine that I wrote\nand maintain.\n\nIn the beginning, I hacked up function mechanisms to return multiple value\nsets and columns. Then PostgreSQL aded \"setof\" and it is was cool. Then it\nwas able to return a set of rows, which was ev... |
[
{
"msg_contents": "If that ID is the only thing you use to access that data, why not just store\nit in a flat file with fixed-length records? seek() (or your language's\nequivalent) is usually fast. \n\nIf you need to drive that from within PostgreSQL, you would need an\nuntrusted language to read the file, but... |
[
{
"msg_contents": "> Lately, I have been thinking that a cool form of index would be some\nsort\n> of \"persistent reference\" index. Like the old ISAM days of yore, a\nfixed\n> number could point you right to the row that you want. I'm not sure if\nthe\n> \"persistent reference\" is a specific auto numbering c... |
[
{
"msg_contents": "Is there a way to bypass the optimizer and to specify a plan to be executed?\n\nSomething like:\n\n\n\nLimit \n -> Nested Loop \n -> Nested Loop \n -> Seq Scan on tab00 t00 \n -> Index Scan using tab03_pkey on tab03 t03 \n Index Cond: (\"outer\".id = t03.id... |
[
{
"msg_contents": "Hi,\n\n--- a little off topic ---\n\nI want to do something so informative like the map that was in the\nprev main developer page.\n\nCan anyone point me about what tool to use? and maybe some guidance on this?\n\nthanks,\nJaime Casanova\n",
"msg_date": "Fri, 11 Feb 2005 00:45:44 -0500",
... |
[
{
"msg_contents": "Hi list,\n\nI am working with Postgres 8.0.0 final on a SuSE Linux 9.2 box. I tried\n\ndropdb ax20050206\nWARNING: could not remove database directory \n\"/export/home/postgres/data2/base/115101837\"\nDROP DATABASE\n\n(maybe there was still some vacuum activity on the database...)\n\na) Ther... |
[
{
"msg_contents": "hello all\n\nas i have mentioned this before, as part of my academic project\ni am required to make modifications in postgresql.\n\nOne of my tasks includes allowing plpgsql functions to pass parameters by\nreference. For this purpose I have been digging in the code for plpgsql \nand postgres... |
[
{
"msg_contents": "\n> I asked the question how do you get a record without going through an\n> index, the answer was CTID, which unfortunately changes when the row is\n> updated.\n\nThe ctid is a physical location of the row. On update a new tuple is written \nin a new location, that is why the ctid changes. T... |
[
{
"msg_contents": "I was asked to give a lectures (a whole day) about PostgreSQL in Moscow, Russia.\nSince I'm new to such kind of activity I'n looking for any presentations,\ncurlliculum for starting.\n\n \tRegards,\n \t\tOleg\n_____________________________________________________________\nOleg Bartunov, sci.r... |
[
{
"msg_contents": "Is VACUUM in 8.x supposed to be overestimating reltuples by 50% of\nthe number of dead tuples?\n\nCREATE TABLE foo (x integer);\nINSERT INTO foo SELECT * FROM generate_series(1, 1000);\nUPDATE foo SET x = x;\nUPDATE foo SET x = x;\nUPDATE foo SET x = x;\nVACUUM foo;\nSELECT relpages, reltuple... |
[
{
"msg_contents": "Hi all,\n\nWhen I try to input a unicode caracter which code is greater than \nU+20000, phpPgAdmin returns the following error message :\n\nERROR: Unicode characters greater than or equal to 0x10000 are not \nsupported\n\nCould someone fix this problem ?\nIf yes, would you please tell me whe... |
[
{
"msg_contents": "Hi All\n\nI have a quite strange problem with RETURN NEXT statement.\nI have a big table with 500 millions of rows running on Postgres 8.0. \n Table \"public.usno\"\n Column | Type | Modifiers \n--------+--------+-----------\n ra | real | \n dec | real | \n bmag | real | ... |
[
{
"msg_contents": "Jeroen van Iddekinge <iddekingej@lycos.com> writes:\n> begin;\n> create table a0(a bigint);\n\n> than login for a second session\n> begin\n> create table a0(a bigint)\n\n> postgres block nows in session 2\n\n> when session 1 is commited the following error appears in session 2\n\n> duplicate... |
[
{
"msg_contents": "Hi,\n\nI noticed that there where strange symbols in the error message when I \ncan't connect to a database.\nThis happends in PHP and pgsql and in the 7.4.x and 8.0 version of \npostgesql\n\nin pqStrError there is a 'if def' for 'char*' and 'int' return type \nstrerror_r.\nI'm have FC2 lin... |
[
{
"msg_contents": "I'm working on an experimental patch to break up the BufMgrLock along\nthe lines we discussed a few days ago --- in particular, using a clock\nsweep algorithm instead of LRU lists for the buffer replacement strategy.\nI started by writing up some design notes, which are attached for\nreview i... |
[
{
"msg_contents": "I'm currently trying to find a clean way to deal with network-dead \nclients that are in a transaction and holding locks etc.\n\nThe normal \"client closes socket\" case works fine. The scenario I'm \nworried about is when the client machine falls off the network entirely \nfor some reason (e... |
[
{
"msg_contents": "Hello all\n\nIs it possible to get the oid of a function on the basis of its name?.\n\nThe scenario which i am currently facing is that i have the function name, now\ni want search the pg_proc system catalog on the basis of the function\nname and retrieve its Oid.\n\nAnother confusion which i... |
[
{
"msg_contents": "Just got reminded...\n\nIs there a way for a C function to determine the name of the schema in which is was created?\n\n... John\n\n\n",
"msg_date": "Mon, 14 Feb 2005 19:32:15 +1100",
"msg_from": "\"John Hansen\" <john@geeknet.com.au>",
"msg_from_op": true,
"msg_subject": "Sch... |
[
{
"msg_contents": "Hi folks,\n\nI ran into big trouble - it seems that my DB is lost.\n\n\"select * from pg_database\" gives me 0 rows, but I still can connect to\ndatabases with \\c and even select from tables there, although they're also\nnot visible with \\dt.\n\nAfter asking the guys in the #postgresql@irc.... |
[
{
"msg_contents": "Beautiful, just what I was looking for.\n\nThnx,\n\nJohn\n\n> -----Original Message-----\n> From: Tom Lane [mailto:tgl@sss.pgh.pa.us] \n> Sent: Tuesday, February 15, 2005 6:31 AM\n> To: Michael Fuhr\n> Cc: John Hansen; pgsql-hackers@postgresql.org\n> Subject: Re: [HACKERS] Schema name of func... |
[
{
"msg_contents": "FYI, core has discussed the pending IBM ARC patent and the usage of\nthose ideas in 8.0.\n\nTom has found a 2Q cache algorithm that predates the ARC patent and is\nvery similar to ARC. The major difference is that it doesn't auto-size\nthe ARC sub-buffers. \n\nCore believes it is best to ba... |
[
{
"msg_contents": "Should our multi-byte encoding be referred to as UTF8 or Unicode?\nI know UTF8 is a type of unicode but do we need to rename anything from\nUnicode to UTF8?\n\nSomeone asked me via private email.\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.... |
[
{
"msg_contents": "I will be on Boston for Linuxworld from Tuesday through Thursday. I\nwill read email only occasionally.\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 359-1001\n + If your life is a hard drive, | 13 Roberts Ro... |
[
{
"msg_contents": "\n> And the user maintenance of updating those hints for every release of\n> PostgreSQL as we improve the database engine.\n\nI don't think so. Basically an optimizer hint simply raises or lowers the cost \nof an index, mandates a certain join order, allows or disallows a seq scan ...\nImho i... |
[
{
"msg_contents": "I've tested the performance of 8.0.1 at my dual-boot notebook (Linux and\nWindows XP).\n\nI installed 8.0.1 for Linux and Windows XP, and run pgbench -c 1 -t 1000\nUnder Linux (kernel 2.6.10) I got about 800 tps, and under Windows XP -\nabout 20-24 tps.\n\nNext I switched off virtual memory u... |
[
{
"msg_contents": "> I've tested the performance of 8.0.1 at my dual-boot notebook \n> (Linux and Windows XP).\n> \n> I installed 8.0.1 for Linux and Windows XP, and run pgbench \n> -c 1 -t 1000 Under Linux (kernel 2.6.10) I got about 800 tps, \n> and under Windows XP - about 20-24 tps.\n> \n> Next I switched o... |
[
{
"msg_contents": ">>> I've tested the performance of 8.0.1 at my dual-boot notebook\n>>> (Linux and Windows XP).\n>>>\n>>> I installed 8.0.1 for Linux and Windows XP, and run pgbench\n>>> -c 1 -t 1000 Under Linux (kernel 2.6.10) I got about 800 tps,\n>>> and under Windows XP - about 20-24 tps.\n>>>\n>>> Next I... |
[
{
"msg_contents": "I was at Linux world Tuesday, it was pretty good. I was in the \"org\"\npavilion, where the \"real\" Linux resides. The corporate people were on the\nother side of the room. (There was a divider where the rest rooms and\nelevators were.)\n\nI say that this was where the \"real\" linux resides... |
[
{
"msg_contents": "> What kind of performance difference can be expected between \n> Linux and Windows? \n\nNot really known, as the native win version is a bit too new for that.\n\nExpect linux performance to be better, though, that much is sure.\n\n//Magnus\n",
"msg_date": "Wed, 16 Feb 2005 13:05:03 +0100... |
[
{
"msg_contents": "> Question 1: Is your writeback cache really disabled in Linux, on the\n> harddrive? Windows fsync will *write through the disk write cache* if\n> the driver is properly implemented. AFAIK, on Linux if write cache is\n> enabled on the drive, fsync will only get into the cache.\n> 800tps sound... |
[
{
"msg_contents": "Hello everyone,\n\nI'm working on a project which makes extensive use of the PostgreSQL table inheritance features and am interested in fixing some bugs I have come across. To my knowledge, triggers, primary keys, index, foreign key and unique constraints are not inherited. I am more than wil... |
[
{
"msg_contents": "Hi,\n\nlooking for the way how to increase performance at Windows XP box, I found\nthe parameters\n\n#fsync = true # turns forced synchronization on or off\n#wal_sync_method = fsync # the default varies across platforms:\n # fsync, fdat... |
[
{
"msg_contents": "> looking for the way how to increase performance at Windows XP box, I\nfound\n> the parameters\n> \n> #fsync = true # turns forced synchronization on or\noff\n> #wal_sync_method = fsync # the default varies across platforms:\n> # fsyn... |
[
{
"msg_contents": "Hi there,\n\nis't possible to make table to be inherited from another table in case\nboth tables already exist. I tried to insert record to pg_inherits,\nbut it doesn't helped.\n\nopenfts=# select * from pg_inherits;\n inhrelid | inhparent | inhseqno \n----------+-----------+----------\n ... |
[
{
"msg_contents": "\"Sergey E. Koposov\" <math@sai.msu.ru> writes:\n> But concerning to the added columns, I can say that before the column errbox\n> didn't, and I revealed the problems began after adding it (together with\n> index creation on that column).\n\nDoes starting a new backend session make the proble... |
[
{
"msg_contents": "Found on a certain other DBMS' mailing list:\n\n * InnoDB: When MySQL/InnoDB is compiled on Mac OS X 10.2 or earlier,\n detect the operating system version at run time and use the\n `fcntl()' file flush method on Mac OS X versions 10.3 and later.\n Apple had disabled `fsync()' i... |
[
{
"msg_contents": "Some SQL constructs will be satisfied before all rows of a set has been \nexamined. I'm thinking of for instance:\n\nEXISTS(SELECT * FROM y WHERE y.a > 0)\n\nIf the first row of collection y fulfills the WHERE predicate, there's \nno reason to continue perusing the rest of the rows. Now, what... |
[
{
"msg_contents": "> ... c) would be very bad since it \n> doesn't give me any chance to release the resources that \n> where used in order to produce the rows.\n\nYou are supposed to free resources used to produce the rows before srf_return_next();\nThe actual rows are pfree()'d by pg. (an dso are any other pa... |
[
{
"msg_contents": "> Hi,\n> \n> looking for the way how to increase performance at Windows XP \n> box, I found the parameters\n> \n> #fsync = true # turns forced \n> synchronization on or off\n> #wal_sync_method = fsync # the default varies across platforms:\n> ... |
[
{
"msg_contents": "\n> Would there be any value in incrementing by 2 for index accesses and 1 \n> for seq-scans/vacuums? Actually, it should probably be a ratio based on \n> random_page_cost shouldn't it?\n\nWhat happens with very small hot tables that are only a few pages and thus have \nno index defined.\n\nI... |
[
{
"msg_contents": "> Things worth experimenting with (these are all untested, so please\n> report any successes):\n> 1) Try reformatting with a cluster size of 8Kb (the pg page size), if\n> you can.\n\nWhat about recompiling pg with a 4k block size. Win32 file cluster\nsizes and memory allocation units are bot... |
[
{
"msg_contents": ">>So by all means turn off fsync if you want the performance gain *and* \n>>you accept the risk. But if you do, don't come crying later that your \n>>data has been lost or corrupted.\n>\n>>(the results are interesting, though - with fsync off Windows \n>and Linux \n>>are in the same performan... |
[
{
"msg_contents": ">> This is what we have discovered. AFAIK, all other major databases or\n>> other similar apps (like exchange or AD) all open files with\n>> FILE_FLAG_WRITE_THROUGH and do *not* use fsync. It might \n>give noticably\n>> better performance with an O_DIRECT style WAL logging at \n>least. But I'... |
[
{
"msg_contents": ">> Things worth experimenting with (these are all untested, so please\n>> report any successes):\n>> 1) Try reformatting with a cluster size of 8Kb (the pg page size), if\n>> you can.\n>> 2) Disable the last access time (like noatime on linux). \"fsutil\n>> behavior set disablelastaccess 1\"\... |
[
{
"msg_contents": ">>Doesn't Windows support O_SYNC (or even better O_DSYNC) flag \n>to open()?\n>>That should be the Posixy spelling of FILE_FLAG_WRITE_THROUGH, if the\n>>latter means what I suppose it does.\n>\n>They should, but someone said it didn't work. I haven't \n>followed up on it, though, so it is qui... |
[
{
"msg_contents": "> >Doesn't Windows support O_SYNC (or even better O_DSYNC) flag to\nopen()?\n> >That should be the Posixy spelling of FILE_FLAG_WRITE_THROUGH, if the\n> >latter means what I suppose it does.\n> \n> They should, but someone said it didn't work. I haven't followed up on\n> it, though, so it is ... |
[
{
"msg_contents": "> > WinXP fsync = true 20-28 tps\n> > WinXP fsync = false 600 tps\n> > Linux fsync = true 800 tps\n> > Linux fsync = false 980 tps\n> \n> Wow, that's terrible on Windows. If there's a solution, it'd be nice\nto\n> backport it...\n> \n\nthere is. I just rigged up a te... |
[
{
"msg_contents": ">> > WinXP fsync = true 20-28 tps\n>> > WinXP fsync = false 600 tps\n>> > Linux fsync = true 800 tps\n>> > Linux fsync = false 980 tps\n>> \n>> Wow, that's terrible on Windows. If there's a solution, it'd be nice\n>to\n>> backport it...\n>> \n>\n>there is. I just rig... |
[
{
"msg_contents": "> One point that I no longer recall the reasoning behind is that xlog.c\n> doesn't think O_SYNC is a preferable default over fsync. We'd\ncertainly\n> want to hack xlog.c to change its mind about that, at least on\nWindows;\n> assuming that the FILE_FLAG way is indeed faster.\nI also confirm... |
[
{
"msg_contents": "> \"Magnus Hagander\" <mha@sollentuna.net> writes:\n> > Tom, if you look at all the requirements of FILE_FLAG_NO_BUFFERING\non\n> >\nhttp://msdn.microsoft.com/library/default.asp?url=/library/en-us/fileio/\n> > base/createfile.asp, can you say offhand if the WAL code fulfills\nthem?\n> \n> If... |
[
{
"msg_contents": ">After multiple runs on different blocksizes( a few anomalous results\n>aside), I didn't see a whole lot of difference between\n>FILE_FLAG_NO_BUFFERING being on or off for writing performance.\n>However, with NO_BUFFERING set, the file is not *read* cached at all.\n>While the performance is o... |
[
{
"msg_contents": "> \"Magnus Hagander\" <mha@sollentuna.net> writes:\n> > Is there actually a reason why we don't use O_DIRECT on Unix?\n> \n> Portability, or rather the complete lack of it. Stuff that isn't in\nthe\n> Single Unix Spec is a hard sell.\n\nWell, how about this (ok, maybe I'm way out in left fie... |
[
{
"msg_contents": "Just wondering after this discussion:\n\nIs transaction wraparound limited to a database or to an installation ?\ni.e. can heavy traffic in one db affect another db in the same installation ?\n\n",
"msg_date": "Fri, 18 Feb 2005 10:53:32 +0100",
"msg_from": "=?ISO-8859-1?Q?J=FCrgen_Cap... |
[
{
"msg_contents": "\n> One point that I no longer recall the reasoning behind is that xlog.c\n> doesn't think O_SYNC is a preferable default over fsync. \n\nFor larger (>8k) transactions O_SYNC|O_DIRECT is only good with the recent\npending patch to group WAL writes together. The fsync method gives the OS a \n... |
[
{
"msg_contents": "\n\n\n-----Original Message-----\nFrom: pgsql-hackers-owner@postgresql.org on behalf of Oliver Jowett\nSent: Fri 2/18/2005 11:27 AM\nTo: Karel Zak\nCc: List pgsql-hackers\nSubject: Re: [HACKERS] UTF8 or Unicode\n \nKarel Zak wrote:\n\n>> Yes, I think we should fix it and remove UNICODE and WI... |
[
{
"msg_contents": "I'm communicating with psql via a pipe stream. This works pretty well, \nbut one problem I have is trying to cancel an operation. If I send a \nsigint, psql dies. In looking at the source I gather this is because it \nassumes I'm in non-interactive mode (pset.notty is true). I was \nwondering... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.