Thanks for all the feedback.  I was thinking that a compile-time
option might be good to avoid eating up the precious getopt space for
things that are rarely used, but this is pretty unanimous.  :)

On Mar 16, 5:38 am, Trond Norbye <[email protected]> wrote:
> Dustin wrote:
> >   Chris Goffinet removed another magic number by making the TCP listen
> > backlog runtime configurable here:
>
> >  http://github.com/lenn0x/memcached/commit/7d010a855ecccf4960fce4ffa4b...
>
> >   Before this change, memcached had a hard-coded backlog of 1024.
> > It's possible to exceed 1024 with a connection burst from a large
> > number of clients at which point some clients will see connection
> > refused.  Some such large installations have pushed this up to 4096 or
> > so, while others might appreciate saving some kernel memory by keeping
> > it low and having fewer bursty connection bringups.
>
> >   Does anyone have an opinion on this?  My only question is whether it
> > should be a compile-time option or a runtime option.
>
> I would prefer to keep the compile-time options at a minimum, because it
> makes it difficult to provide a binary version....
>
> Btw. If you apply this patch please add the -b to the list of options
> in  http://code.google.com/p/memcached/issues/detail?id=28
>
> Cheers,
>
> Trond

Reply via email to