Some of that reported unusable is actually usable.

VxFS doesn't make many inodes in advance, it allocates & creates them
as-needed.
UFS, by contrast, creates inodes at a fixed ratio to storage, and the
space those inodes consume truly is unusable.

But, if VxFS reports that space as free, then as people start to create
files, VxFS start to make inodes, and support get complaints "Hey, I'm
not able to use all the space you said was free, it doesn't behave like
UFS did." So, VxFS makes a conservative estimate of how much space the
inodes are likely to take up. If you make one single 150GB file, I
believe you'll find the inode overhead is much smaller than the current
estimate; if you make a billion zero-size files, then VxFS estimates of
overhead will be too low.

If you want to measure the log size separately, you can get that from
the original mkfs command or it's echo; you can actually adjust the log
size online, but it's a performance/overhead tradeoff. VxFS has
generally been increasing the default log size on the belief that most
customers will find that particular tradeoff beneficial, but you're free
to experiment and change it.

Regards,
Scott

> -----Original Message-----
> From: [EMAIL PROTECTED] 
> [mailto:[EMAIL PROTECTED] On Behalf 
> Of Prasanta Dash
> Sent: Friday, August 10, 2007 11:22 AM
> To: veritas-vx@mailman.eng.auburn.edu
> Subject: Re: [Veritas-vx] Vxfs space missing on new volumes
> 
> Hello,
> 
>  > I just created 3 new volumes and cannot figure out why 
> there is so much  > space missing when I do df -k on my 
> Solaris 8 server:
>  > /dev/vx/dsk/gemdg-san/gem5
>  >                      157227000  104776 147302092     1%
> 
> I believe as a non root user you can use what is listed as available. 
> Difference
> bewteen toal and used is avilable to superuser. This way 
> superuser can do some operations on the filesystem even after 
> it is reported as full. Also per my understanding once you 
> reach this point it will show you the actual available capacity.
> 
> [EMAIL PROTECTED]:/ # vxassist -g local make test 10m 
> [EMAIL PROTECTED]:/ # mkfs -F vxfs /dev/vx/rdsk/local/test
>     version 7 layout
>     20480 sectors, 10240 blocks of size 1024, log size 1024 blocks
>     largefiles supported
> [EMAIL PROTECTED]:/ # mount -F vxfs 
> /dev/vx/dsk/local/test /mnt [EMAIL PROTECTED]:/ # df -k /mnt
> Filesystem            kbytes    used   avail capacity  Mounted on
> /dev/vx/dsk/local/test 10240    2141    7600    22%    /mnt
> [EMAIL PROTECTED]:/ # dd if=/dev/zero bs=4k count=2000 
> of=/mnt/xyz
> 2000+0 records in
> 2000+0 records out
> [EMAIL PROTECTED]:/ # df -k /mnt
> Filesystem            kbytes    used   avail capacity  Mounted on
> /dev/vx/dsk/local/test 10240   10141      99   100%    /mnt
> [EMAIL PROTECTED]:/ # l /mnt
> 
>         Notice that initially it reported u=2141, a=7600 t-(u+a)=500.
>         When I created a 8000 block file, now it u=10141 
> (2141 metadata,
>         8000 data), a=99 and u+a == t .
> _______________________________________________
> Veritas-vx maillist  -  Veritas-vx@mailman.eng.auburn.edu 
> http://mailman.eng.auburn.edu/mailman/listinfo/veritas-vx
> 

_______________________________________________
Veritas-vx maillist  -  Veritas-vx@mailman.eng.auburn.edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-vx

Reply via email to