This is a discussion on FET_BUF_SIZE and The Perl DBI within the Informix forums, part of the Database Server Software category; --> Hi, I have a Perl script, using the DBI and DBD::Informix, and unloading data to disk. To reduce network ...
| |||||||
| FAQ | Members List | Calendar | Search | Today's Posts | Mark Forums Read |
| ||||
| Hi, I have a Perl script, using the DBI and DBD::Informix, and unloading data to disk. To reduce network traffic, the FET_BUF_SIZE has been set to 16k, but the script then dies with "Memory fault". I've set the "LongReadLen" attribute of the db handle, but it does not seem to make any difference. When FET_BUF_SIZE is unset, the script runs fine. Is this a Perl (specific) problem? uname -sr is "Linux 2.4.10-4GB" perl's v is "5.6.1" with Compile-time options: USE_LARGE_FILES |
| ||||
| tvilliers wrote: > I have a Perl script, using the DBI and DBD::Informix, and unloading data > to disk. To reduce network traffic, the FET_BUF_SIZE has been set to 16k, > but the script then dies with "Memory fault". I've set the "LongReadLen" > attribute of the db handle, but it does not seem to make any difference. > When FET_BUF_SIZE is unset, the script runs fine. > > Is this a Perl (specific) problem? > > uname -sr is "Linux 2.4.10-4GB" > perl's v is "5.6.1" with Compile-time options: USE_LARGE_FILES I've responded on the dbi-users@perl.org mailing list. There's likely a bug in CSDK - it shouldn't be crashing. There's also a bug in the use of "16k" (which was enclosed in quotes in the posting to dbi-users!) - you should write 16384. I suspect that the crash is because CSDK is using a null pointer because the conversion of 16k is failing, or is producing the answer 16 and that is just too small. No proof; just a semi-plausible conjecture. -- Jonathan Leffler #include <disclaimer.h> Email: jleffler@earthlink.net, jleffler@us.ibm.com Guardian of DBD::Informix v2003.04 -- http://dbi.perl.org/ |