Last modified: 2010-05-15 15:38:09 UTC
Hi, I'm trying to set up a mirror of the spanish wikipedia. I've installed mediawiki 1.5rc4 as I wanted to import a XML dump (http://download.wikimedia.org/wikipedia/es/pages_full.xml.gz) using the importDump.php script. First time I discovered that my php installation couln't open >2GB files, so I recompiled it with the -D_FILE_OFFSET_BITS=64 following the instructions on http://es2.php.net/manual/en/function.fopen.php#37791. Apparently php can now open such files, so i retryed the import, but now it always stops at the same point, no matter if I cat the .gz to the script or pass the uncompressed file directly to the script: leonov:/var/www/wiki/maintenance# php importDump.phps /var/cache/wikipedia_es/pages_current.xml 100 (154.975063164 pages/sec 154.975063164 revs/sec) . . . 2000 (124.46568431 pages/sec 124.46568431 revs/sec) leonov:/var/www/wiki/maintenance# I' tried the script linked from the "Bug 2979: import script is broken (or dump corrupted?)" (http://bugzilla.wikimedia.org/show_bug.cgi?id=2979) with the same result, also tryed the workaround mentioned in "Bug 3182: Not enough memory for ImportDump.php" (http://bugzilla.wikimedia.org/show_bug.cgi?id=3182). The php process seems not to grow in memory size anyways. The server is a dual Xeon 3.0Ghz with 4GB RAM. Is there any other version of the script I should try? Thanx
*** This bug has been marked as a duplicate of 3473 ***
