Dries Feys
Member
Hi all,
We're using the stateless appserver since a few years succesfully. However, we still have 1 important question for the management of the appservers.
We're running several persistent procedures with many internal procedures in each of them. Some of these persistent procedures create some dynamic query and can pass over the handle of the temp-table to the client side, using a stub procedure. This approach works very fine and very fast for our company. On the other hand, we also use many SDO-components on that particular appserver. The SDO may contain calculated fields which are the result of one of the (many) persistant internal procedures.
But... it can always happen that we 'mismade' a query with a bad index, or that a user demands some huge query on a SDO, again with a bad index.
We easily can discover on what table the bad query is running on (by using some database performance programs), but until now we haven't found any way to discover a way to know which procedure is running that particular query. If we simple kill the specific appserver agent, the problem is resolved, but we never really can discover the exact cause of the problem. Is there anyone who knows a way to do some procedure monitoring without having to change the logging level? (The logfiles cause a full file system if we let them on verbose for a few days...)
We're using the stateless appserver since a few years succesfully. However, we still have 1 important question for the management of the appservers.
We're running several persistent procedures with many internal procedures in each of them. Some of these persistent procedures create some dynamic query and can pass over the handle of the temp-table to the client side, using a stub procedure. This approach works very fine and very fast for our company. On the other hand, we also use many SDO-components on that particular appserver. The SDO may contain calculated fields which are the result of one of the (many) persistant internal procedures.
But... it can always happen that we 'mismade' a query with a bad index, or that a user demands some huge query on a SDO, again with a bad index.
We easily can discover on what table the bad query is running on (by using some database performance programs), but until now we haven't found any way to discover a way to know which procedure is running that particular query. If we simple kill the specific appserver agent, the problem is resolved, but we never really can discover the exact cause of the problem. Is there anyone who knows a way to do some procedure monitoring without having to change the logging level? (The logfiles cause a full file system if we let them on verbose for a few days...)