Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 11 May 2005 14:02:44 -0400 (EDT)
From:      Jerry McAllister <jerrymc@clunix.cl.msu.edu>
To:        freebsd.org@donnacha.com
Cc:        freebsd-questions@freebsd.org
Subject:   Re: I need further HDD advice before submitting order.
Message-ID:  <200505111802.j4BI2ihS009417@clunix.cl.msu.edu>
In-Reply-To: <4282357F.40901@donnacha.com>

next in thread | previous in thread | raw e-mail | index | archive | help
> 
> Jerry, thank you so much.
> 
> > Make a file system out of the 'unallocated' space now even if you don't
> > decide on a mount point or use until later.
> 
> Does it matter what I call that file system i.e. can I change it easily 
> later on from the command line?  Can I just call the free space on the 
> 1st HDD /freea, and on the 2nd /freeb ?  And should I be dividing each 
> into more than just one file system?

You don't need to mount it until you decide what to do with it
so, no, it doesn't need to be called anything.

> > The system uses swap space for its business regardless of how much ram
> > you have - even if it is not forced to swap to have enough space.
> > It uses it for paging as well as swap too.
> 
> Thanks, didn't think of that.  Am I right to assume that paging isn't 
> massively intensive?  And that it's still a good idea to have all the 
> swap on the 1st HDD (as opposed to split between the two) because it 
> will be balanced out anyway by tmp, home and var on the 2nd HDD.

I wouldn't call it massively intensive - unless you do run low on
memory.    I tend to prefer splitting it between drives, but I suspect
it doesn't really matter a lot unless you get to the point of running
low on memory.   

One more thing.   Do not use the socalled 'dangerously dedicated' disk setup.
Just make one regular slice on each disk that uses all of the disk.  
Then partition those slices.   You won't notice the difference in 
amount of available disk and you might save yourself some headaces
later on.
 
> 
>  > That should be plenty for /tmp
> 
> Okay, so, my provisional setup at the moment is:
> 
> P4 2.8Mhz FSB 533, 2GB RAM
> 
> 80GB HDD IDE:
> / = 1GB
> /usr = 10GB
> /local = 10GB
> Swap = 4GB
> /freea = 50GB
> 
> 200GB HDD IDE:
> 
> /tmp = 2GB (is that enough?)
> /home = 28GB
> /var = 100GB (will inclube the forum databases etc)
> /freeb = 70GB
> 
> How does that sound?  Again, I would be very grateful for any and all 
> advice and opinions.

Well, I would do it differently, but I am only responsible for my
machines and not yours.  

I would put all of the standard partitions on the smaller 
disk (/, /tmp, /usr, /var, and even /home if you don't need any 
more than 28GB, plus a chunk of the swap)  
I would not make a separate /usr/local.

I would divide up the large disk between some swap and all the rest in
one big partition for spill.   Then as things grew and began to outstrip
their space on the smaller drive, I would just move the the specific
directories on to the larger drive and make symlinks.  That way I don't
have to pre-guess how much any part is going to grow.  I also would not
have to mess with growfs or anything like that.

So, my more likely allocation would be something like:

On the 80 GB drive
  /root    1 GB
  swap     2.5 GB
  /tmp     2 GB
  /usr     5 GB
  /var     5 GB
  /home    64GB  (eg. all the rest, see size notes below)

On the 200 GB drive
  swap     2.5 GB
  /spill   197 GB (eg. all the rest, see size notes below)

Things that might be likely to need moving to /spill would 
be /var/db, /var/mail, /var/spool, /usr/ports, /usr/local.

Keep in mind that a nominal 80 GB drive will look like 74.5 GB to the 
system because of the difference in the manufacturer's definition of 
a GB (80 billion in decimal) and the system's definition in which a 
GB is 2^30 or 1,073,741,824.   Similarly, the 200 GB drive will really
be 186.26 GB in system style GBs.  That gets written off as a matter of
semantics. But it is really a marketing ploy similar to the gas stations 
naming their prices with the .9 on the end as in 205.9 cents per gallon 
rather than just making it 206 cents so people will think 205 rather 
than 206.

But, to add to this, once you get your 74.5 GB drive and start to set
it up, when you newfs the partitions to make file systems out of them, 
you will lose some more to space it needs for superblocks and pionters.
And then when the file system is mounted, it will reserve 0.08 (eg 8 per 
cent) to deal with system issues.  

So, in reality, your advertised 80 GB drive will yield about 65 GB for 
regular use.

The advertised 200 GB drive will yield about 163 GB after newfs and 
system reserve.

So, with that in mind, you may also want to rethink the size of your
partitions.  Allocating 1 GB will yield about 830 MB filesystem, etc.
Or, to get a 5 GB filesystem, you would need to allocate about 6 GB
for the partition before newfs and system reserve.

One more thing to think about is the possibility of creating a raid (mirror)
for critical things.  That would give you some security in case of a
disk crash.   In that case you would need to put your critical stuff
on the smaller drive and carve out an identical size piece on the large
one to be its mirror partner.   Check out vinum(8).  That, of necessity,
would make your allocation scheme look quite different.
 
////jerry

> 
> Thanks,
> 
> Donnacha
> 
> 
> > ////jerry
> > 
> > 
> > 
> >>I decided not to use GPT because, although it sounds great, it seems a 
> >>little complicated for a newbie like me.
> >>
> >>Apologies for seeking your help once again, I just need to get this 
> >>straight before submitting the order, I would be very grateful for any 
> >>and all advice.
> >>
> >>Thanks,
> >>
> >>Donnacha
> >>
> >>
> >>_______________________________________________
> >>freebsd-questions@freebsd.org mailing list
> >>http://lists.freebsd.org/mailman/listinfo/freebsd-questions
> >>To unsubscribe, send any mail to "freebsd-questions-unsubscribe@freebsd.org"
> >>
> > 
> > 
> 
> 
> 



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200505111802.j4BI2ihS009417>