Page MenuHomePhabricator

importDump.php: Oracle DB error "quoted string not properly terminated" after several hours
Closed, DeclinedPublic

Description

Author: dbq343

Description:
Error Message ORA-01756: quoted string not properly terminated is thrown after several hours of errorfree processing of the Wikipedia XML Dump.

The last DB query was

SELECT /* LinkCache::addLinkObj 127.0.0.1 */ * FROM (SELECT page_id,page_len,page_is_redirect,page_latest FROM /*Q*/PAGE WHERE page_namespace = '0' AND page_title = 'Diné_Bahaneʼ' ) WHERE rownum >= (1 + 0) AND rownum < (1 + 1 + 0)


Version: 1.19
Severity: blocker
OS: Linux
Platform: PC
Whiteboard: aklapper-moreinfo

Details

Reference
bz38793

Event Timeline

bzimport raised the priority of this task from to Medium.Nov 22 2014, 1:08 AM
bzimport set Reference to bz38793.
bzimport added a subscriber: Unknown Object (MLST).

The query looks good. I wonder if he doesn't like the quoted utf-8 characters.

dbq343 wrote:

After switching the DB Charset to UTF-8 i cannot complete the setup process:

Creating main page with default content... failed
Information

Could not insert main page: ORA-02091: transaction rolled back ORA-02291: integrity constraint (MEDIAWIKI.REVISION_FK2) violated - parent key not found

Steve: Which exact MediaWiki version is this about, and is this still a problem?

freak wrote:

I'm closing this one ... the database charset setting is only the internal default (for internal sessions i.e triggers, jobs, aq) how the data is stored and retrieved depends on client NLS settings (which in MW is hardcoded to AL32UTF8).

There could be a php/oci problem ... i've heard (never actualy hit that problem) that sometimes the driver can't handle the fact that the database and client are not using the same NLS setting, but i couldn't reproduce it.

In case you (Steve) still have there this problem, please free to reopen it or contact me directly.