MYSQL conf

[jaymc]

I am in the process of re tweaking my mysql conf

Rather than ask a specific question firsrt, I just want to get a general idea of what to indentify a problem

For example, I read that:

query_cache_size= 32MB for every 1GB of RAM

Now, if I have 3GB of RAM, that would mean 96MB right?

I actually have it set to 512MB. I have never reverted back, due to the fact I never go into SWAP. With this said, if there is no swap, increasing to 512 is not causing a problem thus can only be better than having it set to 96MB

Or am I missing something?

Same applies for

key_buffer = 128MB for every 1GB of RAM

I read that those 2 our really important, so again I have assigned 956MB to the key buffer and still the memory threshold seems to cope as I never go into swap


Is this a good way to go about it, or will this cause problems. When my site gets busy, problems are popping up.

Should I revise my my.cnf methods or is swap = 10k a reason to turn a blind eye to the points I have mentioned above

THANKS!

[cpanelkenneth]

If your caches are too large, then it will take an abnormally long time for the data in them to expire. This means SQL statements will produce incorrect results.

[jaymc]

If your caches are too large, then it will take an abnormally long time for the data in them to expire. This means SQL statements will produce incorrect results.

A side from incorrect data being given, could it have a reverse effect having cache so large?

[cpanelkenneth]

What do you mean by "reverse effect" ?

[jaymc]

What do you mean by "reverse effect" ?

Where as typically the more memory you can give to something

But, giving to much may effect other features or services

I mean, at the moment there is no swap, so Im not running out, but is it causing secret problems that arnt blatantly obvious

[cpanelkenneth]

If you are asking:

By giving a lot, or too much, RAM to the MySQL process, will other processes suffer?

The answer is: it depends.

It depends upon typicall system load, amount of RAM total in the server, which Operating System is used, which kernel is used, other processes running, etc.

By configuring the caches to be overly large, MySQL is also put under extra burden as it must track the data in the cache, which adds to its own management overhead.

[jaymc]

Cool, I have heard it isnt too good anyway

I think I might cut it to 256 MB