L
Libor Laubacher
Guest
>> Any idea how to avoid this? (2nd time in the last few weeks. previously no problems. The only thing I believe that has happened is the splitting up of sql and 4gl clients on to different servers). << Spliting up the clients to go to their own broker was a good idea, it should be like this. On how to avoid it - best is to find the query causing this (unless there is someone or something killing the process), but narrowing down the query can be long and painful process. If the queries you are running haven't changed or new ones added, I'd check indexes (idxcheck) to make sure there is no corruption there and also would set -Ma 1 and -Ma -1 for the SQL broker, so should this happen again, the client won't take up others with it - as Paul suggested a client could have "just" crashed without doing anything disruptive on its own, but since it crashed the server, it killed off other clients with it and one of that of in MTX. Actually, I recall a similar issue in past, there were SQL crashes taking the database down, we never got to 100% pinpoint the reason of the crashes (bug suspected, but upgrade wasn't possible at that time). after setting -Ma & -Mi 1, the SQL client occassionally still kept crashing, but since its alone there, it never took the db anymore down with it.
Continue reading...
Continue reading...