July 2009 - users-list

[MonetDB-users] merovingian / php
by dariuszs 18 Jul '09

18 Jul '09

Hi, If merovingian starts database on demand after a crash then how would I know which port to use for connection in php script, it errors out for port 50000? $db = monetdb_connect("sql", "localhost", 50000,"monetdb", "monetdb_password");

2 1

[MonetDB-users] production server
by dariuszs 18 Jul '09

18 Jul '09

Hi, Does anybody use MonetDB in production environment? Thanks. Dariusz.

3 4

[MonetDB-users] varchar type with copy into + misleading error message
by Franck Routier 17 Jul '09

17 Jul '09

Hi, using copy into, I run into a problem where a string does not fit into my varchar column. So the first point is that the error message is quite misleading, as it states 'field 4 not inserted, expecting type str'. Well, it is a string, but the length is too long... Using insert into, the error message will be 'SQLException:str_cast:value too long for type (var)char(100)' which is much more helpful. The second point is that my string is 99 characters long, and utf-8 with accented characters. So its length is less than 100 characters but more than 100 bytes. In postgresql, where my data come from, varchar(100) means 100 characters, whatever encoding the database happen to be in. Oracle gives the choise with varchar2(100 char) or varchar2(100 byte) and will default to NLS_LENGTH_SEMANTICS (which in turn is BYTE by default). I guess in monetdb, it means 100 bytes... Am I right stating that, and is it the expected behaviour or a bug ? Franck

2 2

[MonetDB-users] Merovingian Neighbour discovery
by Ross Bates 17 Jul '09

17 Jul '09

Trying to get Merovingian to act as a front end to MonetDB cluster. When I start the Mero service this shows up in the log. 2009-07-16 11:28:10 ERR merovingian[2276]: HELO message: error while sending message: Network is unreachable 2009-07-16 11:28:10 ERR merovingian[2276]: ANNC message: error while sending message: Network is unreachable I can connect back and forth between the two Mero/Monet instances using mclient --lsql, but for some reason the instances of Mero can't discover eachother. I think Mero is using UDP which could be contributing to the issue. Any tips for troubleshooting Mero connectivity over UDP? Related to this, is there anyway I can just tell one instance of Mero about the other without auto discovery and proxy that way? Thanks! -Ross

2 2

[MonetDB-users] Jdbc errors : no prepared statement
by Franck Routier 15 Jul '09

15 Jul '09

Hi, trying to populate my database using jdbc, I encounter the following error : java.sql.SQLException: no prepared statement with the given id at nl.cwi.monetdb.jdbc.MonetConnection $ResponseList.executeQuery(MonetConnection.java:2219) at nl.cwi.monetdb.jdbc.MonetConnection $ResponseList.processQuery(MonetConnection.java:1978) at nl.cwi.monetdb.jdbc.MonetConnection.commit(MonetConnection.java:348) The error comes at a seemingly random time, depending on the run, I could insert 22000 rows, then 9800, etc. I am using the latest published version (MaySP1). Is this a know problem ? Franck

3 6

[MonetDB-users] Inner join & nulls performance question
by Alex Bo. 15 Jul '09

15 Jul '09

Hi, I want to do a join between tables on a column that contains nulls. What should I expect performance-wise when running SELECT ... FROM... INNER JOIN... ON (a.f1 = b.f1 OR (a.f1 is null AND b.f1 is null)) compared to a normal join: SELECT ... FROM... INNER JOIN... ON (a.f1 = b.f1) Thanks. -- View this message in context: http://www.nabble.com/Inner-join---nulls-performance-question-tp24496517p24… Sent from the monetdb-users mailing list archive at Nabble.com.

1 0

[MonetDB-users] Select * ... LIMIT takes long time
by Alex Bo. 15 Jul '09

15 Jul '09

Hi, I am running the May2009 SP on Windows. I have populated a table with 40M rows and 20 columns. It takes 'SELECT * FROM table LIMIT 10' 2.5 minutes to return, in which memory and cpu consumption seems to be at a standstill. Why does it take so long? -- View this message in context: http://www.nabble.com/Select-*-...-LIMIT--takes-long-time-tp24493160p244931… Sent from the monetdb-users mailing list archive at Nabble.com.

2 1

[MonetDB-users] server messages
by dariuszs 14 Jul '09

14 Jul '09

Hi, What do these mean? >#GDKmmap(394264576) fails, try to free up space [memory in use=36011928,virtual memory in use=511214485504] #GDKmmap(394264576) fails, try to free up space [memory in use=36011952,virtual memory in use=511214485504] #GDKmmap(394264576) result [mem=34705504,vm=286633623552] #GDKmmap: recovery ok. Continuing.. #GDKmmap(394264576) result [mem=34379040,vm=228672143360] #GDKmmap: recovery ok. Continuing.. Thanks. Dariusz.

2 1

[MonetDB-users] server crash/lib_sql.so.2.30.0
by dariuszs 14 Jul '09

14 Jul '09

Hi, What does this mean? Jul 14 11:41:05 monetdb kernel: [78349.487006] mserver5[18645]: segfault at 18 i p 00007f2ed339d62d sp 00007f2ed2cd3dc0 error 4<6>mserver5[18651]: segfault at 18 ip 00007f2ed339d62d sp 00007f2ed20cddc0 error 4 in lib_sql.so.2.30.0[7f2ed32da0 00+103000] Thanks. Dariusz.

2 3

[MonetDB-users] Data Loss after Restarting MonetDB
by JLee 14 Jul '09

14 Jul '09

Hi: We are running into an interesting problem that we can’t seem to resolve. We have a large number of xml files that we are parsing and loading to MonetDB Feb2009-SP2 sql server running on Redhat with 40gigs of memory. Basically, everything looks fine while the files are being loaded. But, once the server is shutdown and restarted (after the data load has been finished), all the data is simply gone! We have done a dump and restored the db before restarting the db, then loaded more data on top of it; and the data that was restored is intact. But, the newly loaded data is simply gone. It doesn’t seem to matter how small or large the data set being loaded is (tried as small as 1.5 gig of xml files resulting in 1.2M rows). Which leads me to believe our data is not persistent! (a minor problem.. J) We also noticed that as our parser (java app using JDBC with hibernate) is loading the data, that the memory use continues to grow and never shrinks after the data load is complete. So, it seems like the all the data being loaded is being kept in the memory; but never becomes persistent. We tried waiting 4 hours after loading 1.2M rows, thinking that there may be a lag in syncing to the disk (which seem odd, in case of a sudden server failure); but the data simply is gone. Can you provide some assistance in troubleshooting this problem? Thanks. Best regards, H

1 0