Re: Hard-limits for kern.maxproc
Re: Hard-limits for kern.maxproc
- Subject: Re: Hard-limits for kern.maxproc
- From: Rand Childs <email@hidden>
- Date: Thu, 05 Feb 2009 14:48:36 -0600
I have to agree with Nathan and take exception to some of the comments
that Terry wrote about OS X's configuration design.
Many modern operating systems don't have hard limits compiled into the
kernel for many of these kernel resources but instead allocate these
resource dynamically. AIX and later versions of Solaris are capable
of doing this both in 32 and 64 bit versions. Configurable maximum
parameters are usually provided as a tool for the OS administrator to
control the kernel resources on the available hardware, not to limit
the kernel's potential resources, and therefore limit the control of
the system administrator. The administrator should control the
kernel, not the other way around.
In my opinion, adjusting internal kernel parameters and recompiling a
kernel to increase kernel resource should not be necessary in a modern
operating system that is expected to provide production services such
as those advertised by OS X Server.
Personally I ran up against the kernel limit of the maximum number of
SYS V message queues while porting and testing a Unix application I
was working on. As far as I could tell there was no easy way to
increase the number of message queues without serious modification to
the Darwin code. Instead I wrote a message queue library replacement,
implementing the minimum features that I needed using semaphores and
shared memory. AIX, Solaris and other vendors Unix implementations,
and Linux can be configured to run the application without recompiling
the kernel and without writing a message queue library to replace the
one provided with the operating system. OS X was the odd man out.
One would hope that Apple's Mac OS X designers and developers will
eventually provide the same modern kernel resource allocation and
configuration features that are already provided in other Unix/Linux
operating systems if Apple hopes to be able to sell Mac OS X Server
into any kind of serious production environment.
Rand
On Feb 5, 2009, at 12:51 PM, Nathan wrote:
Terry, I understand the intent you are trying to express (the default
settings need to work well for the general case) but I disagree with
your assertion that having such extremely low limits (max processes in
particular, but I suspect other settings as well) serves that goal.
I'm running out of political and motivational steam to pursue this
issue, as I've worked around the most pressing problems we were
experiencing internally. I've canceled most of our plans regarding
the Xserve and OS X server, as I no longer have any confidence in
Apple's ability to deliver on their marketed server features or
capacity. The problem seems to run deeper than just bugs.
I will continue to keep an eye out for shifts in the landscape for OS
X server. I hope to be proven wrong in the future, and find that OS X
Server delivers on its promises.
~ Nathan
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Darwin-kernel mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Darwin-kernel mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden