We have some strange results on a Redhat 4.0 server.
When executing a query, the server directly going to ‘copying to tmp table’ and the server is finished after 1-2 minutes.
Same software, same db, same my.cnf on a Windows OR FreeBSD server, the ‘copying to tmp table’ does not occur.
Query time 8-10 seconds! both FreeBSD and Windows servers are older machines with less ram Our linux machine has 4 gig.
Analysing the query and DB.
OOPS: missing some indexes. Adding them:
Cool!! our linux machine is now finished in 8 seconds. (copying to tmp table shows up for only 2-3 seconds now)
Windows and FreeBSD now needs more time than before!!!
This is complete weird.
Does the mysql optimizer do some strange things under redhat?
Don’t know where to start looking now.
linux machine: dell 2950 2x dualcore xeon 4gig mem
Freebsd: dell dualcore 1,6 2 gig mem
Windows: Amd 1800 512mb
WHY does the ‘copying to tmp table’ not occur on the other machines.