Just curious about something. But I don't know the technical vocabulary, so please forgive this verbose (and confusing?) explanation.
I think it essentially has do with user queues. I have a Rev.cgi script on the On-Rev server that works in this 3-step sequence: (STEP 1) uses SELECT to retrieve data from a MySQL table (STEP 2) does stuff based on that retrieved data in order to alter it (STEP 3) uses UPDATE to put that now-altered data back into the same MySQL table My question is about data mishaps that could happen during STEP 2 -- specially if it happens to take a long time (a few seconds?) -- should it happen to come about that a new user calls a Rev.cgi which is still working on a previous user's call. The chaos sequence I have in mind is this: USER 1 : STEP 1 --> Rev.cgi retrieves data from SQL table at time X ( = tData) USER 1 : STEP 2 --> Rev.cgi works with tData to make tData_User1 (but no UPDATE yet) ... but during USER 1 : STEP 2 ... USER 2 : STEP 1 --> Rev.cgi retrieves data from SQL table at time X+1 (still = tData) USER 2 : STEP 2 --> Rev.cgi works on tData to make tData_User2 (but no UPDATE yet) USER 1 : STEP 3 --> updates SQL table with tData.User1 USER 2 : STEP 3 --> updates SQL table with tData.User2 In this chaos scenario of mine -- if indeed this is how things actually happen between servers, which I'm hoping it's not -- the problem is the state of tData at USER 2 : STEP2. It should not be tData but actually tData_User1. So, is there something in-built in SQL or Rev.cgi that handles queues in such a way that prevents this kind of queue chaos, whether it's 2 or 2 million users calling in at the "same" time Or do I have to take care of this myself somehow, by, say, proper scripting? Thank you. -- Nicolas Cueto _______________________________________________ use-revolution mailing list [email protected] Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-revolution
