logical volumes under 3.3

Chris Wagner jwag at moose.asd.sgi.com
Tue Sep 18 03:00:47 AEST 1990


In article <13693 at hydra.gatech.EDU>, robert at shangri-la.gatech.edu
(Robert Viduya) writes:
> We got 3.3 earlier this week and I've been reading over the manuals trying
> to plan an install later this month.  One of our systems has 4 780M disks
> and I was thinking of combining parts of all four into one giant logical
> volume and use it for user file space (one of the biggest pains in the
> rear on our other systems has been balancing disk utilitization across
> multiple file systems dedicated to users).  Well, as I was working out
> the details, I realized that the resulting logical volume would end up
> being well over 2G long.  Furthermore, we may be adding more disk in the
> future and I'd like to be able to just add them to the logical volume.
> What happens if it goes over 4G?  With a 32-bit int/long, we can get a
> maximum of 2G (signed) or 4G (unsigned).  The 3.3 manuals say you can keep
> adding disk partitions to logical partitions and doesn't mention any limits
> at all.  Somehow I would be very surprised if this were true (pleasantly
> surprised, but surprised nonetheless).
> 
> The man page for statfs(2) seems to indicate that it could handle a
> file system that was 1024G long (using 512-byte blocks), but the man
> page for stat(2) indicates that the max file size is 2G.  If this is
> truely the case, I'd be satisfied; we probably aren't going to have very
> many files over a gigabyte.
> 
> Has anyone who's gotten 3.3 played with this stuff yet?  Is there a limit
> on how large logical volumes can be?
> 
> 			robert
Well, there is some missing dicumentation! In 3.3 the
maximum file size is still 2G (cause of lseek, etc). The maximum
file SYSTEM size is 8G due currently to on-disk space limitations.

Chris Wagner



More information about the Comp.sys.sgi mailing list