Re: message.log: lock table overflow

> 20:21:42 lock table overflow - userid ..., sessionid ... [...] > LOCKS of $INFORMIXDIR/etc/$ONCONFIG is set to 350000 > > Does the 'lock table overflow' message mean the server need more LOCKS? > Can I solve the problem by increasing LOCKS value ? > Any response to solve the problem is appreciated ! You can also check the "ovlock" parameter as displayed in "onstat -p" to see a count of how many times you have run out of locks since reboot or counter reset. Use "onstat -k" to see who is holding locks. Even better, build a routine which queries the SMI tables and presents you with a customized report. The OnLine Administrator's Guide provides a map of the supported tables, and you are welcome to use the others at the risk of them changing without notice. It does sound like you need more locks. You can increase them to the max value of 8,000,000 or change the application/user load to use fewer locks at a time. TFM says that locks are cheap, increasing the amount of shared memory as you add more. Be aware that locking is not simple, and that due to referential constraints, indexes, shared and promotable locks and such that even a simple transaction can lock many rows. ______________________________________________________ Clem Akins (aka clema@informix.com) Informix Software, Inc (Standard disclaimers apply) International Technical Support Last seen: Lunching with the Executives in Singapore