I have a major problem with my node table crashing every time cron runs. I have to repair the table after every run, which is clearly not ideal in any sense. it also crashes randomly when authors try to edit their posts.

Errors look like this:

Incorrect key file for table './node.MYI'; try to repair it query: DELETE FROM node WHERE nid = 23908 in /var/www/html/includes/database.mysql.inc on line 66.
	Incorrect key file for table './node.MYI'; try to repair it query: DELETE FROM node WHERE nid = 23909 in /var/www/html/includes/database.mysql.inc on line 66.
Incorrect key file for table './node.MYI'; try to repair it query: UPDATE node SET title = 'blah', body = 'blah' in /var/www/html/includes/database.mysql.inc on line 66.
Unknown column 'grant_create' in 'where clause' query: SELECT COUNT(*) FROM node_access WHERE (nid = 0 OR nid = 20695) AND CONCAT(realm, gid) IN ('all0') AND grant_create = 1 in /var/www/html/includes/database.mysql.inc on line 66.

i've granted all privs to the database account and flushed, so it isn't a locking issue. I've repaired the whole database and did a clean restart. mysql version is 5.0.24. no matter what happens, it seems whenever anyone tries to change or delete a node, the whole thing locks up and we're back to the repair.

i have found these bug reports for mysql:
bugs.mysql.com/22384
lists.mysql.com/bugs/303
bugs.mysql.com/bug.php?id=4198

which all sound very similar, but dbadmin is not my expertise and i am on a paid host, so i don't think i can do much to the mysql server.

so i suspect that somewhere, the various node tables and node revision tables are out of synch with each other. does anyone know how to correct the key file and can advise? we have to get this resolved before we can start our long over due upgrade process.

thanks to any angels who can help.

Comments

agentrickard’s picture

We need more details. Specifically, Drupal version and MySQL version.

Drupal 4.6.x does not have a separate {node_revision} table. There is also no 'grant_create' column in the {node_access} table.

Also, what installed modules (contrib) are you using? I'd like to know what cron process is calling for node deletion.

My hunch is that you're using an incompatible module that is causing these problems.

--
http://ken.blufftontoday.com/
http://new.savannahnow.com/user/2
Search first, ask good questions later.

darthcheeta’s picture

my hosting provider has confirmed that the delete crash is a bug in the version of mysql they have and doesn't have to do with drupal.

the grant_column seems to be left over from when we were experimenting with taxonomy_access, so we've dealt with that.

but, it looks like we need to find a new host. aggregator, aggregator2, blogging and forums are the core of our site. can't live without them. although we've set ag2 and the others to never delete, cron is weeding out old aggregator content with each run. it can only be set to max of 16 weeks.