I am working with an application generating a pretty big database and generating some very big indexes (all sorts of - including FULL TEXT) to one of the table in particular. It has now lasted for more than an hour with that single table. It has not been possible to query anything about that table for so long. INDEX file grows steadily and is now about 5 times as big as the DATA file (tables are MyISAM) I must assume that the application LOCKS the table for READ while building the indexes (I cannot see the SQL as it is a java application).

Also when starting SQLyog and connecting to the host, 'autocomplete' gets stuck, as it cannot query the table either. If I click the 'powertools' menu rebuilding of tags stops/aborts and the 'rebuilt tags' menu item is disabled.

Now, TABLE LOCKS lasting - say some minutes or even more - probably are quite common, I guess.

Shouldn't 'autocomplete have a time-out period for each DATABASE and TABLE? Say 5 sec's .. 10 sec's. If not it has been able to query the DATABASE/TABLE for that amount of time, it should be skipped and it should proceed to next DATABASE/TABLE?

And actually, in this particular situation there is no way to stop the running query either. No 'red cross' icon, and program is not responsive. So the request for a time-out when accessing a LOCKED table is not only for autocomplete!

Thanks for the info. This is definitely a good idea. I will add this in the TO-DO list.

We can basically implement a user defined timeout error for all kinds of Autocomplete query.



The red cross action should work. Once you execute a query and it stops, pressing the red cross should break the query and reconnect again. Is this not happening or am I missing something?

The red cross simply does not show up in this situation.

Maybe the order-of-statements in the code?
First show read cross .. next send query ?