Making a dynamically-linked root
Alexey Dokuchaev
danfe at nsu.ru
Tue Jun 3 09:27:38 PDT 2003
On Tue, Jun 03, 2003 at 09:14:52AM -0700, Matthew Dillon wrote:
>
> :On Mon, Jun 02, 2003 at 02:25:43PM -0700, Matthew Dillon wrote:
> :> start!). Running certain daemon startups in the background might yield
> :> a significant overall improvement in startup times.
> :>
> :> e.g. instead of running 'sshd' you would run sshd in a subshell, aka
> :> (sshd &), so the RC script can continue on with the next thing without
> :> having to wait for sshd to fault-in from disk. Same goes for sendmail
> :> and many other daemons.
> :
> :This isn't a definite win. I know in the past it used to actually
> :slow things down: To take your example, having both sshd and sendmail
> :attempting to fault-in from disk in parallel will thrash both the disk
> :and cache far more than sshd and sendmail sequentially faulting in. A
> :very large number of daemons trying to start in parallel will also
> :stress the scheduler.
> :
> :Peter
>
> I'm fairly sure there isn't an issue. Both a hard drive's own
> on-board cache and FreeBSD's clustering and caching code are *very* well
> suited to this sort of parallel initiation. There is certainly no
> scheduler issue. The key advantage here is that you are removing
> serialization that would otherwise cause both cpu cycles and disk
> cycles to be wasted waiting for each other. Take sendmail for example.
> sendmail usually takes upwards of a second to startup due to initial
> DNS lookups that it makes and other things. sshd doesn't start
> instantaniously either, I think due to creating the initial
> session keys.
Methinks that Matt's suggestion of (sshd &) and alikes does indeed
sound very cool, and at least worth of reference implementation and
seeing it in action. As already mentioned by Matt, the bottleneck here
is not in I/O (which is, in fact, really fast nowadays) but in either
network access (in case of sendmail) or (and?) CPU-intensive routines
(speaking of sshd, respectively).
Just my $.02 though, as it all had been already said more or less the
same way by other folks.
./danfe
More information about the freebsd-arch
mailing list