|
Hello,
On Fri, 9 Feb 2001, Joseph Mack wrote:
> Joseph Mack wrote:
>
> With sufficient number of connections, a director could
> start to swap out its tables (is this true?)
IMO, this is not true. LVS uses GFP_ATOMIC kind of allocations
and as I know such allocations can't be swapped out.
> In this case, throughput could slow to a crawl. I presume
> the kernel would have to retrieve parts of the table to find
> the real-server associated with incoming packets. I would
> think in this case it would be better to drop connect
> requests than to accept them.
>
> In earlier verions of LVS, you could set the amount of
> memory for the tables (in bytes). Now you allocate a number
> of hashes, whose size could (in the worst case) grow without limit.
>
> If I have 128M RAM == 1M connections, am I going to
> run out of ports first?
No, we will run out of memory first :) We will not reach
out of ports :) If we run web cluster for example, we use only one
port, 80.
> What would have to happen for LVS to start swapping?
> (ie how much of a problem is this?)
>
> If it's possible for LVS to start the director to swap,
> is there some way to stop this?
You can try with testlvs whether the LVS uses swap. It is easy,
just start the kernel with LILO option mem=8M and with large swap area.
Then we must check whether more than 8MB swap will be used.
> Joe
Regards
--
Julian Anastasov <ja@xxxxxx>
|
