basic partition question

You can change the partition flag from unnassigned to usr and save the partition table without affecting the underlying data. However, if you now attempt to do anything with the partition - i.e. newfs it, at this point you will be overwriting any data that resides there.

First off, check what is actually mounted as a file system and match this up with your parition table.

#df -k | grep '^/' to show mounted partitions ignore /proc since this is a virtual filesystem. As Mark says partition 6 and 7 may be in use but simply marked unassigned.

#swap -l to list swap partitions

If you are running disk management software, the device paths will contain vx (Veritas) or md (Disksuite). For these you need to see how they have been allocated.

# vxdisk list to list disks under Veritas control and
# vxprint -hrt shows how partitions have been assigned.

# metastat to see how Disksuite is layed out

Commonly Veritas uses encapsulation to take over control of the whole disk - so any unassigned space is allocated as free space under Veritas control.

If your system is simply using normal device files /dev/dsk/... then you still have to carry out a couple more checks to see if the space really is free.

If you have a database system running on your server, it may be using raw devices. ls -l /dev/dsk will show a list of links to the underlying device files. cd to this underlying directory and call fuser -f * to see if any processes are attached to the unassigned slices.

Finally your supplier may have used the unused slices to take a back up a copy of your system before shipping it. By attempting to fsck the raw device you can see if it has a filesystem created on it and what it was last mounted as.

If you are unsure, your best bet might be to call the people who set your system up and find out how the system is layed out. This should really have been documented.

SoupDragon (and Mark)

df -k | grep '^/' returns:

/dev/dsk/c1t1d0s0    2058319 1695233  301337    85%    /
/proc                      0       0       0     0%    /proc
/dev/dsk/c1t1d0s3    2058319   69574 1926996     4%    /var
/dev/dsk/c1t1d0s7    7230605  225733 6932566     4%    /export
/dev/dsk/c1t1d0s5    2058319    2076 1994494     1%    /files
/dev/dsk/c1t1d0s4    2058319  399262 1597308    20%    /opt


Suggesting that slice 6 is not mounted as a filesystem.  This is my partition table information:

Part      Tag    Flag     Cylinders         Size            Blocks
  0       root    wm       0 -   412        2.00GB    (413/0/0)     4202688
  1       swap    wu     413 -  1237        4.00GB    (825/0/0)     8395200
  2     backup    wm       0 - 14086       68.35GB    (14087/0/0) 143349312
  3        var    wm    1238 -  1650        2.00GB    (413/0/0)     4202688
  4 unassigned    wm    1651 -  2063        2.00GB    (413/0/0)     4202688
  5 unassigned    wm    2064 -  2476        2.00GB    (413/0/0)     4202688
  6 unassigned    wm    2477 - 12643       49.33GB    (10167/0/0) 103459392
  7 unassigned    wm   12644 - 14086        7.00GB    (1443/0/0)   14683968

swap -l returns:

swapfile             dev  swaplo blocks   free
/dev/dsk/c1t1d0s1   118,1      16 8395184 8395184

ls -l /dev/dsk returns lots of things, the 2 that are relevant to slice 6 are:

lrwxrwxrwx   1 root     root          42 Mar 20 14:14 c0t6d0s6 -> ../../devices/
pci@8,700000/scsi@6/sd@6,0:g

lrwxrwxrwx   1 root     root          70 Mar 20 14:14 c1t1d0s6 -> ../../devices/
pci@8,600000/SUNW,qlc@4/fp@0,0/ssd@w21000004cfb467e7,0:g

I am unsure how to make it to the raw file directory from here, to try the fuser -f command.

Thanks for the help so far - I've already learned a ton of stuff.  I have also put a call in to the manufacturer for assistance.

Thanks again,

Neal

a couple of things to check:
   metadevice configurations:
   /etc/lvm/md.tab   (see if /dev/dsk/c1t1d0s 6/7 are being
                      configured)

   check to see if you have a database installed or not.
   Ie: oracle, look for the ora.tab file.  It should tell
   you if you have a raw partition configured in your
   system.

you did the df -kl (that looks okay)

read the logbook on your server (someone may have documented it before)  I document the hell out of my
servers.

worse case...install another disk drive on the system, they are very inexpensive now-a-days.

Also, try to mount the disk on a partition (DON'T NEWFS it)
if it mounts, it was newfs'ed before.  If it doesn't mount, it is unformated (but still might be a raw partition for Oracle or something).

(it's always best to be cautious than to kill years worth of data :)

Are there entries in /etc/vfstab for these partitions?

If you have Informix as a database on this system look the are aor raw disk that /dev/online_root points to.

Follow /dev/rdsk/<controller><target><disk>slice6/7 to see if that particular slice is owned by a particular user or group that may identify a use for it.

D.

No comment has been added lately, so it's time to clean up this TA.
I will leave a recommendation in the Cleanup topic area that this question is:

PAQ and refund

Please leave any comments here within the next four days.

PLEASE DO NOT ACCEPT THIS COMMENT AS AN ANSWER!

liddler
EE Cleanup Volunteer