You are not logged in.
Pages: 1
we have similar usage like this:
https://synopse.info/forum/viewtopic.php?id=4712
many TSQLRestServerDB instance hosted in one httpserver,
we do this to preform "Each Client Has It's Own Database", and to practice the "Sharding"
this approach really simplify the project's architecture and coding.
we have 1000 TSQLRestServerDB s hold in each Httpserver ,yes you see it right , 1000+
by test we fill two table of each database with One Million rows ,then each db file in disk is about 200M
we load all the 1000 databases, the project's whole RAM use is just around 300M, so we think with todays machine ablility ,
this will work just fine .
and we run 10+ process on each machine to get maxium use of the machine power.
if this work ok ,it will save a lot of money!!
after reading the linked post above , we afriad this approach may cause some other issue, like backup\ cache\ master\slave etc
any thought?
Offline
Each SQlite3 instance has its own cache size.
You can reduce the default cache size of each database, so that it will use less main memory, but rely on the file system cache.
Of course, I would rather compile into 64-bit, so that you may have more than 2GB for the executable.
Offline
Pages: 1