Dual Core Xeon / i386 install w/ more than 4gb of RAM

[Kevin K]

Thank you all for your suggestions. I have been trying to push to move to amd64 architecture for all the reasons you all stated. For the record, we tested PAE on one machine, booted the kernel w/ nextboot and it crashed about 15 minutes later. I will consider configuring a dump device to analyze the kernel dumps, but for now we reverted to the original i386 kernel and are likely going to scrap the PAE idea and move to amd64.

This was a management decision (obviously) and the people who originally built this box (long before I was there), did not have enough experience or foresight. i was hoping for alternative suggestions to reduce downtime of these boxes, such as recompiling amd64 manually instead of a fresh install.

These boxes are just Apache, Mysql, PHP type boxes. Nothing exotic or fancy.



Thanks again for your suggestions. I am trying my best to relay the reasoning and rock-solid logic ;)




-----Original Message-----
From: owner-freebsd-stable at freebsd.org [mailto:owner-freebsd-stable at freebsd.org] On Behalf Of Ivan Voras
Sent: Wednesday, February 20, 2008 11:35 AM
To: freebsd-stable at freebsd.org
Subject: Re: Dual Core Xeon / i386 install w/ more than 4gb of RAM

Tom Samplonius wrote:

>   Is PAE really that stable?  I thought it was fairly unpolished, mainly because PAE is seen as a weak kludge implemented by Intel because they all thought we would all be using Itanium's by now.  Intel reversed their folly pretty quickly, adopted the x86-64 extensions as-is from AMD, and pushed them onto every piece of silicon they make.

Architecturally, it's a nasty kludge. As far as stability on FreeBSD is concerned, my only machine under PAE with 4 GB RAM (without PAE it would use a bit over 3 GB) is very solid on 6-STABLE.

>   I also really don't know how anyone would properly use 16GB of RAM under PAE anyways?  Each process is going to limited to just under 4GB.  The kernel memory space can't be bigger than 4GB either, so forget about a huge disk cache.

As I understand it, one possible benefit could be to use the memory for disk / file cache. AFAIK the pages are just pages, without distinction where they are mapped, and for example, if you run PostgreSQL, it couldn't use more than 4 GB for its own data (actually closer to 2 GB because of some sysvshm issues) but it will indirectly use the cache.

>   And is there some really stability fear about FreeBSD on x86-64?  Seems just the same as i386.

I agree, FreeBSD on amd64 is very stable.