- users-list - monetdb.org

[MonetDB-users] RHEL 4 x86_64 RPMs
by Henk van Lingen 03 May '06

03 May '06

On Fri, Apr 28, 2006 at 06:44:42PM +0200, Stefan Manegold wrote: > > > RPM build errors: > > > File not found by glob: /var/tmp/MonetDB-root/usr/share/MonetDB/lib/*jar > > Can it be that you don't have Java installed on your system? > > Nevertheless I think this is a bug. > > yes, yes, yes: > > - it should also work with REHL4 on x86_64 --- ate least I don't see any > reason why it should not work, provided the systems installed properly; we > actually build RPMs every night on an Itanium running > "Red Hat Enterprise Linux WS release 4 (Nahant Update 3)". NB: Itanium <> x86_64, right? > - your problem looks indeed as if (a proper) java (>=1.4) is not installed or > at least not found by configure (check the configure output and/or the > config.log) Right, I didn't install java on the build host (didn't know it was a dep for MonetDB) However, now I installed java-1.4.2-gcj-compat-devel and java-1.4.2-gcj-compat. But the problems persist. After ./configure --enable-bits=64 make make dist make rpm I have: asclepius:src/MonetDB-4.10.2 144% grep java config.log Configured with: ../configure --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --enable-shared --enable-threads=posix --disable-checking --with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions --enable-java-awt=gtk --host=x86_64-redhat-linux configure:4068: checking for java configure:4086: found /usr/bin/java configure:4098: result: /usr/bin/java configure:4120: checking for javac configure:4138: found /usr/bin/javac configure:4150: result: /usr/bin/javac configure:4198: checking for javadoc configure:4216: found /usr/bin/javadoc configure:4228: result: /usr/bin/javadoc config.status:955: creating src/mapi/clients/java/Makefile config.status:955: creating src/mapi/examples/java/Makefile ac_cv_path_JAVA=/usr/bin/java ac_cv_path_JAVAC=/usr/bin/javac ac_cv_path_JAVADOC=/usr/bin/javadoc JAVA='/usr/bin/java' JAVAC='/usr/bin/javac' JAVADOC='/usr/bin/javadoc' Do I really need the Sun Java? Cheers, -- Henk van Lingen, Systems & Network Administrator (o- -+ Dept. of Computer Science, Utrecht University. /\ | phone: +31-30-2535278 v_/_ http://henk.vanlingen.net/ http://www.tuxtown.net/netiquette/

3 3

[MonetDB-users] RHEL 3 RPMs
by Henk van Lingen 28 Apr '06

28 Apr '06

Hi, Is there a special reason there are no i686 RPM's anymore like there were until 4.6.0_rc2? I think it's a bit weird there are RPM's for the hobby OS 'FedoraCore' but not for the much used server OS 'RedHat Enterprise 3 (and 4)'? The FedoraCore (which Core?) needs libstdc++.so.6 whereas RHEL 3 only has libstdc++.so.5 ... There also seems to be no SRPM's. Cheers, -- Henk van Lingen, Systems & Network Administrator (o- -+ Dept. of Computer Science, Utrecht University. /\ | phone: +31-30-2535278 v_/_ http://henk.vanlingen.net/ http://www.tuxtown.net/netiquette/

6 14

[MonetDB-users] Issues with concurrent connections using JDBC
by Anderson Supriano 09 Apr '06

09 Apr '06

Hi! I'm working on a research about database performance and I'm running a TPC-W implementation (based on http://www.ece.wisc.edu/~pharm/tpcw.shtml) on some databases (Monet, MySQL, Postgresql). Basically, what is being done is having some emulated browsers generating load on the servlet container (tomcat in my case) that access the database through JDBC. -------------- -------------- ---------------- | Emulated | | Tomcat | | DB | | Browser | -----> | | ------> | | -------------- -------------- ---------------- When running this benchmark with concurrent accesses I'm getting problems. If I try 5 concurrent Emulated Browsers running (and 5 concurrent database connections) everything goes well. The problem starts when I increase the concurrency. With 30 concurrent connections everything goes well for a few minutes and then it stops working. The servlet container stops responding and monet uses low percentage of CPU on this state. When I try to kill tomcat and start again (without killing Monet) it starts working again and stops after a while. Based on this experiment it seems that Monet is still responsive, but the instance of servlet container that was accessing Monet just freezes. I'm trying to investigate it a little more, but since I'm not an expert on MonetDB internals, I'd like to know if there is some known issue like this one or if anyone can have a tip on how to solve this. Since Monet is responsive and the servlet container is freezing, it looks like JDBC driver is doing something that is freezing the container (I used to suspect on application, but since it is working very well with MySQL and Postgresql, it seems to not be the problem). I'll appreciate any help on this subject. Thanks very much for your attention. Best regards, Anderson Supriano

2 2

[MonetDB-users] Commit failure on concurrent transactions
by Anderson Supriano 02 Apr '06

02 Apr '06

Hi! I am running several concurrent transactions on Monet DB using JDBC and I'm getting an error when commiting a transaction: java.sql.SQLException: COMMIT: failed ERROR: COMMIT: transaction is aborted because of concurency conflicts, will ROLLBACK instead at nl.cwi.monetdb.jdbc.MonetConnection$ResponseList.executeQuery(MonetConnection.java:2185) at nl.cwi.monetdb.jdbc.MonetConnection$ResponseList.processQuery(MonetConnection.java:1937) at nl.cwi.monetdb.jdbc.MonetConnection.commit(MonetConnection.java:443) Is there a way that this exception can be handled in order to have all transactions working (even if waiting the last one to be finished)? I am using MonetDB-4.10.0 and SQL-2.10.0. Thanks for your cooperation. Best regards, Anderson Supriano

2 2

[MonetDB-users] JDBC: Not all results returned
by Martin Beck 23 Feb '06

23 Feb '06

Hi, I have a simple SQL script which starts like this: CREATE TABLE people ( pid integer NOT NULL PRIMARY KEY, name varchar(40) NOT NULL DEFAULT '', number varchar(15) NOT NULL UNIQUE DEFAULT '0', email varchar(100) NOT NULL, spid integer NOT NULL DEFAULT '0' ); Afterwards, it inserts 251 People with ascending pids from 0 to 250. I apply it to the MonetDB database either with MapiClient, or JDBCClient. However, if I login with the JDBCClient and do a "select name,number from people;" i only get 250 rows, not 251! But no error is reported. Instead, a "select count(*) from people;" returns 251. if i try the same with the MapiClient, everything works as expected. the problem seems to be with other tables, too, but there the count of returned rows another one, so I expect it has something to with a packet limit, or something? is it a configuration thing or a bug? with 2.8.2 everything worked. if executed from a java program, the statement returns the exception "protocol violation: unexpected line in data block:"

3 8

[MonetDB-users] network setup
by Colin Foss 15 Feb '06

15 Feb '06

How do you configure MonetDB to listen to connections from anywhere besides the localhost? I can make a local connection but as soon as I try to connect from another computer I get an error. I have even tested with netcat (nc) to make sure there wasn't a firewall blocking the connection. Colin Foss

2 1

[MonetDB-users] Mondrian and MonetDB
by Martin Kersten 31 Jan '06

31 Jan '06

LS, I think the combination of the opensource project Mondrian and MonetDB would be a very strong couple. Especially given the history of MonetDB in the area of datamining. However, the resources at the core team are insufficient at this point in time to embark on such a (short term) project. Perhaps someone in our user community is willing to take up the challenge and report on the problems encountered/ results obtained. The core development teams is particularly interested in missing features, interfacing problems, and, ofcourse, performance problems. regards, Martin ----------------------------------------------- Read and respond to this message at: https://sourceforge.net/forum/message.php?msg_id=3551757 By: jhyde I'm always happy to add support to Mondrian for another DBMS. The process is fairly well-worn, since we support several DBs already. It's best if someone familiar with the DBMS goes through the process, then I'll integrate the code and update the doc. The process is: 1. Get the MondrianFoodMartLoader working for your DB. 2. Tweak Mondrian's SQL-generation facilities for the capabilities of your DB. Look at SqlQuery.Dialect and you will see the kinds of things that Mondrian cares about. (If your DB has a fairly expressive SQL, and driver provides the appropriate metadata so Mondrian knows what it's capable of, this step may be trivial.) 3. Add your database name as an option for the <SQL dialect="db"> element. (Code is in Mondrian.xml, and elsewhere.) 4. Get the Mondrian regression suite to run cleanly. 5. Provide some typical connect strings for me to put into mondrian.properties. 6. Give me a contact who is willing to run the regression suite on your DB before each release. ______________________________________________________________________ You are receiving this email because you elected to monitor this forum. To stop monitoring this forum, login to SourceForge.net and visit: https://sourceforge.net/forum/unmonitor.php?forum_id=111375

2 1

[MonetDB-users] MonetDB-SQL 2.10.0 - Status of temorary tables?
by leemarks＠netspace.net.au 31 Jan '06

31 Jan '06

Hi, Are temporary tables currently supported in monetdb-sql? If I do something like, create temporary table foo (abc integer) It fails with a strange error. [Perhaps I'm using invalid syntax?] Thanks in advance, Lee ------------------------------------------------------------ This email was sent from Netspace Webmail: http://www.netspace.net.au

4 8

[MonetDB-users] 8 hours of Photoshop Video Tutorials
by Cole 23 Jan '06

23 Jan '06

1 0

[MonetDB-users] MonetDB/Five slotting: on running M4 next to M5
by Fabian Groffen 25 Dec '05

25 Dec '05

Hi list, Scheduled for next year is the (alpha/beta) release of MonetDB/Five, the successor of the current MonetDB generation. MonetDB/Five is the logical successor to its predecessor in terms of speed and performance. However, MonetDB/Five is not fully backwards compatible with it's predecessor. Instead of using the MIL language for kernel operations, it uses MAL. Although many MIL operations can be easily mapped on MAL operations, a fully automatic (and correct) translation of MIL to MAL is not possible. Directly related to this, is the availability of MonetDB/XQuery, which currently heavily relies on the MIL language. The SQL users of MonetDB will be the only ones that will immediately benefit from the improved MonetDB/Five engine, as SQL is fully functional and there is no difference in language or output. Due to this situation where both M4 (curent MonetDB generation) and M5 (MonetDB/Five generation) have it's own advantages (XQuery support, MIL, MAL, better performance) it is most likely that one would like to run both M4 and M5 engines next to each other to benefit from both. This would currently cause problems when applied out of the box. Given that M4 and M5 would be slotted, and hence installable next to each other, starting it's services would cause problems because both engines try to bind to the same ports by default. To give a short view on the ports in use by default: M4 M5 Mapi 50000 50000 SQL 45123 45123 XQuery 45789 N/A This obviously clases for default MIL/MAL (Mapi) connections as well as SQL connections. XQuery is not an issue for the moment. A few observations: - The actual language spoken over the Mapi server port is engine specific, i.e. MIL for M4 and MAL for M5. This is probably somewhat confusing. For SQL it is clear, and the port specifies which language to use, regardless the engine. - XQuery support is not available for M5 yet, but it should be dealt with as for SQL. Possible scenarios to solve the problems of clashing ports can be: 1. Use different port numbers based on the engine, i.e. use 55123 for SQL on M5 and 45123 on M4. Mapi could become something like 50005 for M5. 2. Build (or extend for instance the DatabasePool) an application that handles transparently the request for a service making use of a combination of port and challenge response to give the client a M4 MIL, M5 MAL, M5 SQL or M4 XQuery connection, using the ports currently known. (Note: this also allows for multi-catalog support for SQL, when the DatabasePool is taken as base.) 3. Ignore the problem and only document it, including how to change the default ports to avoid clashes. (Note: unfortunately we're unable to bind to a certain interface...) 4. A combination between 1 and 2, with the purpose of getting each language it's own port in the system, reducing the problem and possibly disable M4's SQL server. Are end-users interested in/aided by such changes made, or will they figure out themselves easily enough? Just a few thoughts on a problem that will arise in distribution land...

1 0