RE: Locked rows when Online crashes

Richard, In what way does the row "seem to be locked"? Any specific error message? As a guess, I'd say that the update was performed but not committed before the core dump. So, most likely you still have a process out there holding the lock. Try using "onstat -k" to look for the lock or "onstat -g ses" to look for the orphaned process. Whichever is easier for you. Once you find the orphaned session id, you can use "onmode -z <sid>" to kill the process and roll back the transaction. But, make sure to read about "onmode -z" first!!!! Specifically, you don't want to kill it if the process is holding a latch. (onstat -u will show if that's the case). Simple enough?! Actually, if you have more details or if you have specific questions, feel free to write to me directly. --Craig (craig@chemistry.utah.edu) >I am using Online v6.00 with isql and dbaccess on Solaris. I managed to >cause isql to core dump by using a perform screen on a table that had = >been >altered since the form was last compiled - twice :-). The second time, I = >was >in update mode, and the row now seems to be locked. I can query it using = >=20 >perform, but it won't let me update or remove the row. When I was using = >SE >(the good old days - JOKE!) I could simply go into the .dbs directory = >and >type 'clearlok'. With Online, there is no .dbs directory!!!!! >