Hi,
We are using the older 1.8.0.5 version of SMW with MediaWiki 1.20.2, PHP 5.4.9, MySQL 5.5.28. Everything has been working great so far and we are well over 106,000 pages in our system after over 2 years or use in production. Recently we encountered the following database error when running jobs. It seems to occur only a specific type of page invoking a specific template.
A database query syntax error has occurred. The last attempted database query was: "INSERT IGNORE INTO `t6` SELECT t7.smw_id FROM `smw_object_ids` AS t7 WHERE t7.smw_sortkey LIKE 'CID%'" from within function "SMW::executeQueries". Database returned error "1114: The table 't6' is full (localhost)"
Is this related to some default MySQL setting for SMW? We currently have just over 16GB RAM used for MySQL caching and buffers. We have plenty of hard disk space. I'm wondering if there is some simple MySQL setting we need to change to increase teh table size. Any help here would appreciated.
Thanks,
> We are using the older 1.8.0.5 version of SMW with MediaWiki 1.20.2, PHP 5.4.9, MySQL 5.5.28.
Honestly it will be difficult for use to make any suggestions since you are using a no longer maintained code base (SMW 1.8*). It will be toilsome for us to make an appropriate response but we are tracking the issue at [0].
> It seems to occur only a specific type of page invoking a specific template.
Is it possible for you to post the details of your analysis to [0]?
Also, could you try running an isolated stack trace to see whether the queries are caught in a resource loop and link the report to [0]. Limit your pages when running the job because "$wgDebugDumpSql = true" is expensive as it returns each single SQL query executed.
[0] https://github.com/SemanticMediaWiki/SemanticMediaWiki/issues/918
