RE: MySQL Website
It appears to be the web server. I can reach mysql.com just fine in a traceroute, but can't get a HEAD or webpage to come up! traceroute to mysql.com (66.35.250.190), 30 hops max, 38 byte packets 1 cdm-208-180-236-1.cnro.cox-internet.com (208.180.236.1) 18.963 ms 10.260 ms 12.200 ms 2 cdm-208-180-1-50.cnro.cox-internet.com (208.180.1.50) 7.622 ms 9.933 ms 9.904 ms 3 cdm-208-180-1-73.cnro.cox-internet.com (208.180.1.73) 17.948 ms 17.666 ms 14.908 ms 4 dllsbbrc01-gew0402.ma.dl.cox-internet.com (66.76.45.145) 128.870 ms 182.677 ms 91.958 ms 5 dllsdsrc01-gew0303.rd.dl.cox.net (68.1.206.5) 23.685 ms 26.633 ms 22.810 ms 6 dllsbbrc01-pos0101.rd.dl.cox.net (68.1.0.144) 23.805 ms 26.595 ms 27.092 ms 7 12.119.145.125 (12.119.145.125) 79.373 ms 78.874 ms 75.386 ms 8 gbr6-p30.dlstx.ip.att.net (12.123.17.54) 75.101 ms 79.933 ms 74.823 ms 9 tbr2-p013701.dlstx.ip.att.net (12.122.12.89) 82.161 ms 80.284 ms 77.678 ms 10 ggr2-p390.dlstx.ip.att.net (12.123.17.85) 78.322 ms 75.077 ms 81.961 ms 11 dcr2-so-4-0-0.Dallas.savvis.net (208.172.139.225) 76.214 ms 77.886 ms 76.674 ms 12 dcr2-loopback.SantaClara.savvis.net (208.172.146.100) 108.356 ms 105.723 ms 112.343 ms 13 bhr1-pos-0-0.SantaClarasc8.savvis.net (208.172.156.198) 95.535 ms 88.560 ms 84.063 ms 14 csr1-ve243.SantaClarasc8.savvis.net (66.35.194.50) 88.678 ms 86.770 ms 85.408 ms 15 66.35.212.174 (66.35.212.174) 89.425 ms 89.129 ms 98.684 ms 16 mysql.com (66.35.250.190) 87.200 ms 85.178 ms 87.600 ms Thanks, Brad Teale Universal Weather and Aviation, Inc. mailto:[EMAIL PROTECTED] > -Original Message- > From: Peter Burden [mailto:[EMAIL PROTECTED] > Sent: Wednesday, April 21, 2004 7:35 AM > To: Lehman, Jason (Registrar's Office) > Cc: [EMAIL PROTECTED]; [EMAIL PROTECTED] > Subject: Re: MySQL Website > > > Lehman, Jason (Registrar's Office) wrote: > > >I should have been clearer. I can't reach the website. I can get to > >lists.mysql.com with no problem except for the fact that images won't > >pull form www.mysql.com but I definitely come to a grinding > halt when I > >try to reach www.mysql.com. I can't do a tracert because > the university > >has shut that off here. But I guess it is working for everyone else. > > > > > > I'm experiencing similar problems - using both Mozilla and IE. > > 'wget' eventually got the HTML but it took nearly 2 minutes. > The headers don't suggest anything strange. > > > This is also a University site with 'traceroute' disabled and > everything > accessed through a cache. > > www.netcraft.com's site analysis also doesn't suggest > anything untoward. > > >-Original Message- > >From: Rhino [mailto:[EMAIL PROTECTED] > >Sent: Tuesday, April 20, 2004 6:31 PM > >To: Lehman, Jason (Registrar's Office) > >Subject: Re: MySQL Website > > > > > >- Original Message - > > > > > >>From: "Lehman, Jason (Registrar's Office)" <[EMAIL PROTECTED]> > >>To: <[EMAIL PROTECTED]> > >>Sent: Tuesday, April 20, 2004 11:53 AM > >>Subject: MySQL Website > >> > >> > > > > > > > > > >>Does anyone know what is going on with the MySQL website? > >> > >> > > > >It appears to be undergoing a major redesign. The sections > appear to be > >organized differently and the style sheets have also changed. > > > >Or did you have something else in mind? > > > >Rhino > > > > > > > > > > > > > > > > > -- > MySQL General Mailing List > For list archives: http://lists.mysql.com/mysql > To unsubscribe: > http://lists.mysql.com/[EMAIL PROTECTED] > -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
RE: C API
There is a C++ package called OTL (http://otl.sourceforge.net/home.htm). It supports both MySQL through MyODBC, and Oracle. It works great with Oracle applications, but we have not used it with MySQL. Thanks, Brad Teale Universal Weather and Aviation, Inc. mailto:[EMAIL PROTECTED] 713-944-1440 ext. 3623 Arrange things so that a person needs to know nothing, and you'll end up with a person who is capable of nothing. -- K. Brown -Original Message- From: Priyanka Gupta [mailto:[EMAIL PROTECTED] Sent: Monday, October 20, 2003 7:14 PM To: [EMAIL PROTECTED] Subject: C API Is there a way to have a common C API for MySQL and Oracle. I am writing some software that I would like to work with both MYSQL or Oracle as the backend server? priyanka _ Enjoy MSN 8 patented spam control and more with MSN 8 Dial-up Internet Service. Try it FREE for one month! http://join.msn.com/?page=dept/dialup -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED] -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
RE: Ancestry program
It has been a while since I have looked, but I believe the National Genealogical Society has a data model for family tree software. The following links are to the NGS and GEDCOM is the file format standard. I think it should be an easy conversion to a database structure. If you do something that exports the data, it should probably export in the GEDCOM format because that is what most software packages will import. http://www.ngsgenealogy.org/ http://www.gentech.org/ngsgentech/main/Home.asp GEDCOM seems to be the standard file format: http://www.gendex.com/gedcom55/55gctoc.htm Brad -Original Message- From: Dan Greene [mailto:[EMAIL PROTECTED] Sent: Tuesday, October 28, 2003 12:49 PM To: Nitin; [EMAIL PROTECTED] Subject: RE: Ancestry program well... when I do db design, I tend to start with the objects of my system. The one that comes to mind in your case is people. so you'll need a people table. well what are the details of a person? first_name Last_name Middle_name1 Middle_name2 Maiden_name [any other basic bio data] so you'll need those columns Well to keep track of each person, each one will need an ID... id's are usually numbers, so now you add a: person_id field. This field would likely have an auto_increment attribute to help number them for you ok... now that we have people, what else do we need? relationships between them well... in terms of human beings, everyone has one biological mother and one biological father, so we add in mother_id father_id leaving the values of these as null would be equivalent of being 'unknown' and we now have, data-wise, a system that can trace biological heritage, can handle siblings and half-siblings. Other ideas for objects: Marrages - this one would be tricky/interesting, as marrages can change over time, and people can have multiple marrages (although usually not two at a time, unless bigamy is allowed in your user's state/country). Strictly speaking, marrages are not necessary to trace heritage, but are good info... > -->From: Nitin [mailto:[EMAIL PROTECTED] > -->Sent: Monday, October 27, 2003 10:46 PM > -->To: [EMAIL PROTECTED] > -->Subject: Ancestry program > --> > -->Hi all, > --> > -->I'm developing a web based ancestry program. The user > wants it to be > -->static, that means, it isn't for all to use, but his family. Better > to > -->say, it'll contain only his family tree. > --> > -->Now, I cant think of the proper db design, which will help any user > to > -->find his or her relationship with any other person in the tree. > Though, I > -->can design a simple database, where everything will have to be done > -->through queries and scripts, but I want those queries to keep as > simple > -->as possible. > --> > -->Any help will be appreciated, as I'm new to such a problem. > --> > -->Thanx in advance > -->Nitin > > > > -- > MySQL General Mailing List > For list archives: http://lists.mysql.com/mysql > To unsubscribe: > http://lists.mysql.com/[EMAIL PROTECTED] > > -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED] -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
RE: C API
There is a C++ package called OTL (http://otl.sourceforge.net/home.htm). It supports both MySQL through MyODBC, and Oracle. It works great with Oracle applications, but we have not used it with MySQL. Thanks, Brad Teale Universal Weather and Aviation, Inc. mailto:[EMAIL PROTECTED] 713-944-1440 ext. 3623 Arrange things so that a person needs to know nothing, and you'll end up with a person who is capable of nothing. -- K. Brown -Original Message- From: Priyanka Gupta [mailto:[EMAIL PROTECTED] Sent: Monday, October 20, 2003 7:14 PM To: [EMAIL PROTECTED] Subject: C API Is there a way to have a common C API for MySQL and Oracle. I am writing some software that I would like to work with both MYSQL or Oracle as the backend server? priyanka _ Enjoy MSN 8 patented spam control and more with MSN 8 Dial-up Internet Service. Try it FREE for one month! http://join.msn.com/?page=dept/dialup -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED] -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
RE: Adding indexes on large tables
Brendan, We have used ext2, ext3, and reiser for testing purposes, and we have found ext3 to be terribly slow on file read/write operations. If you need a journaling file system, I would go with reiser, otherwise ext2 will be blazingly fast. The other thing I would do is move your DB to another drive like Dan, said. Brad -Original Message- From: Brendan J Sherar [mailto:[EMAIL PROTECTED] Sent: Tuesday, October 07, 2003 6:27 AM To: [EMAIL PROTECTED] Subject: Adding indexes on large tables Greetings to all, and thanks for the excellent resource! I have a question regarding indexing large tables (150M+ rows, 2.6G). The tables in question have a format like this: word_id mediumint unsigned doc_id mediumint unsigned Our indexes are as follows: PRIMARY KEY (word_id, doc_id) INDEX (doc_id) The heart of the question is this: When calling ALTER IGNORE TABLE doc_word ADD PRIMARY KEY(doc_id, word_id), ADD INDEX(doc_id), MySQL proceeds to create a working copy of the table. This process takes over an hour to perform. During this time, disk I/O for the rest of the database (live) reaches a bottleneck, and slows to an unacceptable crawl. Once the copy has been created, MySQL is able to do the actual index build very quickly and efficiently. This process must occur three times daily. A) MySQL creates these temporary tables in the same directory as the original datafile. Is there a way to cause it to use an alternate directory (i.e., on a separate mounted disk)? B) Is there a way to "nice" this process in such a way that the amount of I/O it consumes in performing the copy is restricted to a manageable level so that other requests to the disks can be served in a timely fashion? C) Would abandoning ext3 in favor of ext2 create a substantial difference? D) We're reluctant to upgrade to 4.0 at this point, but were we do so, are there any significant gains in this situation? E) The ALTER TABLE query is performed using perl DBI. Is there a lower level call available which would improve performance? F) Any other ideas or suggestions? The system in question has the following setup: Dual Xeon 2.8, 4G RAM, 2 x 146GB U160 SCSI (10,000 RPM) on RAID 1 (hardware). Redhat 8.0, 2.4.18 kernel, using ext3 fs. MySQL 3.23.56, with myisam tables. Relevant variables: myisam_sort_buffer_size=512M tmp_table_size=128M This is a master, so bin_log is on Thanks in advance for your help, and please keep up the excellent work! Best, Brendan -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED] -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
RE: grant by option on querys
Thank you very much Paul. The order by NULL clause sped the query up from 1.5 minutes to 10 seconds! This is what we were looking for. Thanks, Brad -Original Message- From: Paul DuBois [mailto:[EMAIL PROTECTED] Sent: Thursday, October 02, 2003 9:59 PM To: Brad Teale; '[EMAIL PROTECTED]' Subject: Re: grant by option on querys At 21:07 -0500 10/2/03, Brad Teale wrote: >Hi All, > >I asked earlier about a query being slow, possibly due to MySQL 'Using >temporary; Using filesort' when processing the query. I have done some >testing, and it appears that no matter what data set is used, MySQL always >performs a select with a 'grant by' clause using the temporary and filesort >methods. The only time I could force MySQL into not using these methods >happened when a did a goup by on a column that absolutely contained the same >information. Is this the standard behavior? Is there anyway to get around >this? Is there a MySQL variable I can tweak? Try adding ORDER BY NULL to suppress the implicit sorting that GROUP BY does in MySQL. Of course, that means your results won't be sorted. If you really want them sorted, you might try indexing modelhr, the column you're grouping by. You might try indexing it anyway, in fact. That may give you quicker grouping. > > >My example: >mysql> desc foo; >+--+--+--+-++---+ >| Field| Type | Null | Key | Default| Extra | >+--+--+--+-++---+ >| stn | char(4) | YES | MUL | NULL | | >| modelhr | int(2) | YES | | NULL | | >| f_temp | decimal(6,2) | YES | | NULL | | >| m_temp | decimal(6,2) | YES | | NULL | | >| yearmoda | date | | | -00-00 | | >+--+--+--+-++---+ >5 rows in set (0.00 sec) > >mysql> select * from foo; >+--+-++++ >| stn | modelhr | f_temp | m_temp | yearmoda | >+--+-++++ >| KHOU | 6 | 90.00 | 89.60 | 2003-06-01 | >| KHOU | 6 | 76.00 | 71.60 | 2003-06-01 | >| KHOU | 6 | 75.00 | 73.40 | 2003-06-01 | >| KHOU | 6 | 88.00 | 87.80 | 2003-06-01 | >+--+-++++ >4 rows in set (0.01 sec) > >mysql> explain select stn, modelhr, m_temp from foo group by modelhr; >+---+--+---+--+-+--+--+ - >+ >| table | type | possible_keys | key | key_len | ref | rows | Extra >| >+---+--+---+--+-+--+--+ - >+ >| foo | ALL | NULL | NULL |NULL | NULL | 120 | Using >temporary; Using filesort | >+---+--+---+--+-+--+--+ - >+ >1 row in set (0.01 sec) > >mysql> explain select stn, modelhr, m_temp from foo where stn='KHOU' and >yearmoda = '2003-06-02' group by modelhr; >+---+--+---+--+-+--+--+ - >-+ >| table | type | possible_keys | key | key_len | ref | rows | Extra >| >+---+--+---+--+-+--+--+ - >-+ >| foo | ALL | stn,stn_2 | NULL |NULL | NULL | 90 | Using where; >Using temporary; Using filesort | >+---+--+---+--+-+--+--+ - >-+ >1 row in set (0.05 sec) -- Paul DuBois, Senior Technical Writer Madison, Wisconsin, USA MySQL AB, www.mysql.com Are you MySQL certified? http://www.mysql.com/certification/ -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
grant by option on querys
Hi All, I asked earlier about a query being slow, possibly due to MySQL 'Using temporary; Using filesort' when processing the query. I have done some testing, and it appears that no matter what data set is used, MySQL always performs a select with a 'grant by' clause using the temporary and filesort methods. The only time I could force MySQL into not using these methods happened when a did a goup by on a column that absolutely contained the same information. Is this the standard behavior? Is there anyway to get around this? Is there a MySQL variable I can tweak? My example: mysql> desc foo; +--+--+--+-++---+ | Field| Type | Null | Key | Default| Extra | +--+--+--+-++---+ | stn | char(4) | YES | MUL | NULL | | | modelhr | int(2) | YES | | NULL | | | f_temp | decimal(6,2) | YES | | NULL | | | m_temp | decimal(6,2) | YES | | NULL | | | yearmoda | date | | | -00-00 | | +--+--+--+-++---+ 5 rows in set (0.00 sec) mysql> select * from foo; +--+-++++ | stn | modelhr | f_temp | m_temp | yearmoda | +--+-++++ | KHOU | 6 | 90.00 | 89.60 | 2003-06-01 | | KHOU | 6 | 76.00 | 71.60 | 2003-06-01 | | KHOU | 6 | 75.00 | 73.40 | 2003-06-01 | | KHOU | 6 | 88.00 | 87.80 | 2003-06-01 | +--+-++++ 4 rows in set (0.01 sec) mysql> explain select stn, modelhr, m_temp from foo group by modelhr; +---+--+---+--+-+--+--+- + | table | type | possible_keys | key | key_len | ref | rows | Extra | +---+--+---+--+-+--+--+- + | foo | ALL | NULL | NULL |NULL | NULL | 120 | Using temporary; Using filesort | +---+--+---+--+-+--+--+- + 1 row in set (0.01 sec) mysql> explain select stn, modelhr, m_temp from foo where stn='KHOU' and yearmoda = '2003-06-02' group by modelhr; +---+--+---+--+-+--+--+- -+ | table | type | possible_keys | key | key_len | ref | rows | Extra | +---+--+---+--+-+--+--+- -+ | foo | ALL | stn,stn_2 | NULL |NULL | NULL | 90 | Using where; Using temporary; Using filesort | +---+--+---+--+-+--+--+- -+ 1 row in set (0.05 sec) Thanks, Brad -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Query speed issue
Hello, The problem: I have the following query with is taking upwards of 2 minutes to complete and we need it faster, prefer less than 30 seconds (don't laugh): select modelhr, avg(f.temp-b.temp), avg(abs(f.temp-b.temp)), stddev(f.temp-b.temp), stddev(abs(f.temp-b.temp)), count(f.temp-b.temp) from foo as f, bar as b where f.fyearmoda=b.yearmoda and f.fhr=b.hr and f.stn=b.stn and b.yearmoda >= '2003-01-01' and b.yearmoda <= '2003-01-31' and b.stn='' group by modelhr; When we run explain we get: +---+---+---+-+-+--- ++ | table | type | possible_keys | key | key_len | ref | rows | Extra| +---+---+---+-+-+--- ++ | b | range | PRIMARY,interp_hr | PRIMARY | 7 | NULL | 679 | Using where; Using temporary; Using filesort | | f | ref | stn,fcst | stn | 11 | const,m.yearmoda,m.Hr | 26 | Using where | +---+---+---+-+-+--- +--+--+ Is there a reasonable way to get this query from using temporary and filesort? I tried dumping the data into a temporary table, and the explain ran the same. Also, both MySQL setups perform the same. Any ideasPlease! System/Table Stuff Below - System: dual Xeon 2.4GHz machine with 2G RAM Interconnect: QLogicFC 2200 Disk1: Sun T3 hardware raid5 with Reiserfs (64M cache controller) Disk2: Sun T3 hardware raid0 with Reiserfs (64M cache controller) OS: Red Hat Linux release 8.0 (with qlogicfc module) MySQL1: 4.0.14 - prebuilt MySQL rpm uses Disk1 MySQL2: 4.0.15a - Hand built with MySQL Manual options uses Disk2 The table structures are as follows: CREATE TABLE foo ( yearmoda date NOT NULL default '-00-00', mruntime int(2) NOT NULL default '0', mhr int(3) NOT NULL default '0', fyearmoda date NOT NULL default '-00-00', fhr int(2) NOT NULL default '0', stn varchar(4) NOT NULL default '', temp decimal(6,2) default NULL, ... more but unused data here PRIMARY KEY (yearmoda,mruntime,mhr,stn), KEY stn (stn,fyearmoda,fhr), KEY fcst (stn,yearmoda,mruntime) ) TYPE=MyISAM; CREATE TABLE bar ( stn char(4) NOT NULL default '', hr int(2) NOT NULL default '0', min int(2) NOT NULL default '0', day int(2) NOT NULL default '0', temp decimal(6,2) NOT NULL default '0.00', ... More unused data here yearmoda date NOT NULL default '-00-00', PRIMARY KEY (stn,yearmoda,hr,min), KEY interp_hr (yearmoda,hr,stn) ) TYPE=MyISAM; Table Stats: foo - 38G - Data/18G - Index (326K rows) bar - 24G - Data/14G - Index (35K rows) Thanks, Brad -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
RE: Real-time data warehousing
We have used the predecessor to the OTL for many of our apps and were planning to use the OTL for the new system. I thought the OTL used ODBC to make its connection with databases other than Oracle. I know the OTL supports Oracle natively. Sadly we cannot move to Linux. We managed to get our web servers on Linux, but the big iron will always be Sun here (Company policy). There has been talk of getting Oracle 9i? because Oracle has told us it is much faster, but we are not holding our breath. Thanks, Brad Teale -Original Message- From: walt [mailto:[EMAIL PROTECTED]] Sent: Friday, May 17, 2002 12:27 PM To: Brad Teale Cc: '[EMAIL PROTECTED]' Subject: Re: Real-time data warehousing How are your apps written? We use OTL libaries from http://members.fortunecity.com/skuchin/home.htm which are compiled into our C/C++ code. Moving our apps from oracle to mysql only requires changing 3 or 4 lines per call to the db in the code. Its not odbc "compliant", but still allows our apps to be "farily" portable and fast. We debated rewriting our apps to be ODBC compiant, but figured that was one more layer for bugs and we'd have to switch db platforms 4 times for it to be cost effective. Have you tried Oracle on Linux? We did some testing before Oracle told us the cost of migrating our licence from Oracle8/Solaris to Oracle8i/Linux. We benchmarked our current db server, Sun Ultra single processor 768MB ram, against a 600Mhz 500MB ram Intel/Linux box. The Linux./8i/Intel box smoked our current db server. - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail <[EMAIL PROTECTED]> To unsubscribe, e-mail <[EMAIL PROTECTED]> Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
RE: Real-time data warehousing
I forgot to mention, we have Oracle in-house, and the machine the MySQL database will reside on is a 2 proc Sun box with 1.5G of RAM. The Oracle databases reside on a 16 proc Sun box with 10G of RAM. The decision to go or not go with MySQL is not based on money, it needs to be based on performance. We currently use Oracle in-house for everything, but its speed hasn't been its selling point, and for this application we need lots of speed. That is why we are leaning toward MySQL, but were not sure if it could keep up with the addition of the user community. I had one other question, how much of a performance hit would we take with MySQL if we connected through MyODBC? Thanks again, Brad -Original Message- From: walt [mailto:[EMAIL PROTECTED]] Sent: Friday, May 17, 2002 11:47 AM To: Brad Teale Cc: '[EMAIL PROTECTED]' Subject: Re: Real-time data warehousing Brad, We're in process of evaluating mysql vs our current Oracle 8 system. Importing data is much faster in mysql than oracle according the numbers we're getting. However, from our benchmarking, Oracle seems to be faster on the queries (no writes to db during query time). The table were running our queries against has 46 coulmns and 14 indexes (some columns indexed twice in multi-column indexes). All queries are based on indexed columns. We've also run into some issues trying to delete indexes, 14+ hours before we killed the db and reloaded data, but I may be something stupid. One note on Oracle, $30,000+ for a single processor licence. From our testing, it looks like the bottleneck is disk I/O not processing power. With Oracle, you have better control over which disks your data resides on which lets you balance disk I/O better. However, for $30k, you can buy 10 15,000 rpm drives, stripe them, and then buy another server for replication of data and still have $25K left over. - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail <[EMAIL PROTECTED]> To unsubscribe, e-mail <[EMAIL PROTECTED]> Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
Real-time data warehousing
We are warehousing real-time data. The data is received at up to T1 speeds, and is broken up and stored into the database in approximately 25 different tables. Currently MySQL is doing terrific, we are using MyISAM tables and are storing 24 hours worth of data but we don't have any users and we need to store 72 hours worth of data. Our concern is that when we start letting our users (up to 200 simultaneous) hit the database, we won't be able to keep up with ingesting and serving data with the MyISAM locking scheme. We have tested Oracle and PostgreSQL which fell behind on the ingest. The current production system uses regular ISAM files, but we need to make a certification which requires a relational database. Also, the current production system doesn't have the feature list the new system has. Is there a better database solution or do you think MySQL can handle it? If MySQL can handle it, would we be better off using InnoDB or MyISAM tables? Thanks, Brad - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail <[EMAIL PROTECTED]> To unsubscribe, e-mail <[EMAIL PROTECTED]> Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
RE: Compiling on Solaris
I found the answers to my previous question about MySQL in the manual. Doh! However, when I tried to compile MySQL, I ran into the following error: /bin/sh ../libtool --mode=compile cc -DDEFAULT_CHARSET_HOME="\"/export/home/bteale/mysql-3.23.47\"" -DDATADIR="\"/export/home/bteale/mysql-3.23.47/var\"" -DSHAREDIR="\"/export/home/bteale/mysql-3.23.47/share/mysql\"" -DUNDEF_THREADS_HACK -DDONT_USE_RAID -I./../include -I../include -I./.. -I.. -I..-O -DDBUG_OFF -Xa -fast -native -xstrconst -mt -DHAVE_CURSES_H -I/export/home/bteale/pkgs/mysql-3.23.47/include -DHAVE_RWLOCK_T -c hash.c rm -f .libs/hash.lo cc -DDEFAULT_CHARSET_HOME=\"/export/home/bteale/mysql-3.23.47\" -DDATADIR=\"/export/home/bteale/mysql-3.23.47/var\" -DSHAREDIR=\"/export/home/bteale/mysql-3.23.47/share/mysql\" -DUNDEF_THREADS_HACK -DDONT_USE_RAID -I./../include -I../include -I./.. -I.. -I.. -O -DDBUG_OFF -Xa -fast -native -xstrconst -mt -DHAVE_CURSES_H -I/export/home/bteale/pkgs/mysql-3.23.47/include -DHAVE_RWLOCK_T -c hash.c -KPIC -DPIC cc: Warning: -xarch=native has been explicitly specified, or implicitly specified by a macro option, -xarch=native on this architecture implies -xarch=v8plusa which generates code that does not run on pre UltraSPARC processors "hash.c", line 189: reference to static variable "hash_key" in inline extern function "hash.c", line 229: cannot recover from previous errors cc: acomp failed for hash.c make[2]: *** [hash.lo] Error 1 make[2]: Leaving directory `/export/home/bteale/pkgs/mysql-3.23.47/libmysql' make[1]: *** [all-recursive] Error 1 I looked into hash.c, and the relevant code is as follows: 183: #ifndef _FORTREC_ 184: inline 185: #endif 186: uint rec_hashnr(HASH *hash,const byte *record) 187: { 188: uint length; 189: byte *key=hash_key(hash,record,&length,0); 190: return (*hash->calc_hashnr)(key,length); 191: } To fix the problem I poked around, and ended up commenting out lines 183-185. After this it compiles fine. Will this lead to any problems? Computer Config: SunOS 5.8 sun4u sparc SUNW, Ultra-80 Sun Workshop 6 update 2 5.3 2001/05/15 Thanks, Brad Teale Universal Weather and Aviation, Inc. <mailto:[EMAIL PROTECTED]> - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail <[EMAIL PROTECTED]> To unsubscribe, e-mail <[EMAIL PROTECTED]> Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
Compiling on Solaris
A couple of questions about compiling on Solaris. 1) Are the Sun Workshop 6 compilers supported for MySQL and MySQL++? 1a) Can you use the -native flag without problems? 2) Is the binary distribution compiled with Sun or GNU compilers? Background Info: We are currently trying to ingest 1.5M/sec of weather data into a database, and we have had luck using MySQL 3.23.4x on a Linux 800Mhz machine with 256M of RAM. However, this machine is basically useless for anything else, and it is my desktop. We have several Sun servers with 4+ procs and 4+Gb of RAM, and I thought one would make a good ingest machine. So I would like to compile MySQL and MySQL++ with the Sun compilers to take full advantage of everything the platform has to offer. Any help would be great. Thanks, Brad Teale Universal Weather and Aviation, Inc. <mailto:[EMAIL PROTECTED]> - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail <[EMAIL PROTECTED]> To unsubscribe, e-mail <[EMAIL PROTECTED]> Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
AUTO_INCREMENT question...
I am currently using MySQL to warehouse real-time data, and I have a couple of questions regarding AUTO_INCREMENT columns. OS: Linux/Solaris MySQL version: 3.23.33 Table Types: MYISAM 1) The data is only stored for 24hrs. If I do continuous deletes from the tables, will the AUTO_INCREMENT columns reuse the deleted numbers? 2) I was wondering if the AUTO_INCREMENT columns wrapped back to 0 once they run out of numbers on the top end? 3) If neither of these cases are true, is there a way to simulate number 2? Thanks, Brad Teale Universal Weather and Aviation, Inc. <mailto:[EMAIL PROTECTED]> 713-944-1440 ext. 3623 - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail <[EMAIL PROTECTED]> To unsubscribe, e-mail <[EMAIL PROTECTED]> Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php