Re: Cant reproduce Problem in Lab
Tried a lot of things today - but no success in any way.
I used 5.1, changed everything i could find to latin1, started replication from scratch. It worked for 15minutes, then the old error.
I did put a mysql 5.6 on the slave, tried to change everything to utf8 (5.6 allows the indizes to be larger then 1000) and started the replication from scratch. Started OK, but same old error after an hour.
Then i tried to change 5.6 with latin1 - same error. I also tried to change mysql to cp850 where i was able to - same result.
Changed mysql back to latin1, edited the dump-file of the master and changed all of th UTF8 entrys to latin1 before importing and starting replication anew - same error.
Whatever i do on the slave, replication fails with the same error.
I still cant reproduce it in my lab. Whatever i try there works without problems. I copyed the whole mysql directory from the live master and created the same replication as in the live szenario - with only the slave-port being different. But i can update, insert and delete whatever i want - the slave updates every character like a charm.
I tried to find some infos about the language used in the app on the live master - it seems to be somthing called "Keil uVision" according to the extensions. All i found was that the mysql ODBC-Driver is installed, so my guess is that they use that to access mysql. As said i have no access to the sources of the Program, so i cant tell if there is any "set names" or something like that, but i doubt it very much. To me it seems that there is VERY basic SQL understanding (if i look on the statements replicated). So i dont think there will be issued anything other than a insert/select/delete.