Disksuite questions

finster · September 1, 2003, 1:10am

Is this correct if I have to create 2 partions of 100 Gigs each? (spread accross (6) 36 Gig drives)

AVAILABLE DISK SELECTIONS:
0. c0t0d0 <SUN36G cyl 24620 alt 2 hd 27 sec 107>
/pci@1f,4000/scsi@3/sd@0,0
1. c0t1d0 <SUN36G cyl 24620 alt 2 hd 27 sec 107>
/pci@1f,4000/scsi@3/sd@1,0
2. c1t1d0 <SUN36G cyl 24620 alt 2 hd 27 sec 107>
/pci@1f,4000/scsi@3,1/sd@1,0
3. c1t2d0 <SUN36G cyl 24620 alt 2 hd 27 sec 107>
/pci@1f,4000/scsi@3,1/sd@2,0
4. c1t3d0 <SUN36G cyl 24620 alt 2 hd 27 sec 107>
/pci@1f,4000/scsi@3,1/sd@3,0
5. c1t4d0 <SUN36G cyl 24620 alt 2 hd 27 sec 107>
/pci@1f,4000/scsi@3,1/sd@4,0
6. c1t5d0 <SUN36G cyl 24620 alt 2 hd 27 sec 107>
/pci@1f,4000/scsi@3,1/sd@5,0
7. c1t6d0 <SUN36G cyl 24620 alt 2 hd 27 sec 107>
/pci@1f,4000/scsi@3,1/sd@6,0

1) Format each disk and create 2 - 17 G partitions on each drive(slice 4 and slice 5 for each disk)

2) create initial state database replicas (1 on each disk)

e.g. metadb -a -f c1t1d0 c1t2d0 c1t3d0 c1t4d0 c1t5d0 c1t6d0

3) use disksuite to join slices together making a logical partion of 100gig each?

Current partion of disks 2-7:
Part Tag Flag Cylinders Size Blocks
0 root wm 0 - 90 128.37MB (91/0/0) 262899
1 swap wu 91 - 181 128.37MB (91/0/0) 262899
2 backup wu 0 - 24619 33.92GB (24620/0/0) 71127180
3 unassigned wm 0 0 (0/0/0) 0
4 unassigned wm 0 0 (0/0/0) 0
5 unassigned wm 0 0 (0/0/0) 0
6 usr wm 182 - 24619 33.67GB (24438/0/0) 70601382
7 unassigned wm 0 0 (0/0/0) 0

Will I have to format these and reserve a small slice for database replica ? (This is the way they were purchased.)

I have read the documentation but I am confused as to the state of disk before creationg DB replicas? Should the whole disk(each slice) be unassigned?

Sol 8 installed on disk 0. Having trouble understanding how I should start and how disksuite will handle partions.

please help...

davidg · September 1, 2003, 8:50am

Hi,

I think you should be very carefull now !!!
Please always note down th excisting config of the format command before you take action !!

In this case your c0 disk will almost always be your root-disk. You definitly don't want to change that config !!!

So always start at c1 in this case. As you have all disks on the same controller there is no need to do split mirroring, etc. Just use a whole disk on c1t?d0s2 .Slice 2 is always the complete disk, so no need for re-config.

Please read the docs that can be found on sunsolve.sun.com This is the best site ever for finding Sun solution.
Be proud to do Sun maintenance !! Take your time to read the apropriate docs.

Regs David

finster · September 1, 2003, 10:31am

I am confused as to the state of disk before creationg DB replicas? Should the whole disk(each slice) be unassigned?

What should the disk look like before adding replicas?

davidg · September 1, 2003, 11:15am

Ouch, I think I missed it here. Sorry !

Was reading about databases and stuff not reading well enough.

A replica DB should be always at least 1034 blocks. I don't think you will make it this way.
But isn't it that for root-disks you have a free Veritas license ? I believe so. Sorry to tell, but Solstice is not where you should be up to anymore.
You should just leave it the way it is now and encapsure these disks under Veritas, not bothering about replica databases and stuff anymore

If you still want it, I'dd definitly deep into sunsolve.sun.com or in this case maybe even better docs.sun.com
It has all you ever need. More than just a simple explanory of disksuite.

Regs David

finster · September 1, 2003, 2:01pm

Don't have the option to use veritas...........Thanks.

Should I repartition the 6 drives (say slice 0 = 128mb for database replica), slice 5 - 17 gig, slice 6 17 gig, for all 6 drives? (having 1 database replica for each drive on slice 0). Is this the way the disk is prepared to be use by disksuite?

I need to have 2 100 gig partitions when I am done.

If anyone can help,...it would be appreciated....

Perderabo · September 1, 2003, 5:30pm

I am not real sure what you're asking. But disksuite uses slices as its building blocks. So you are going to need slices.

And all it can do is connect the slices. If you want a 100 gig partition, then you will need some slices that add up to 100 gigs.

As a standard in our company, every disksuite disk must have a database. So that's the only way that I have done it. Make this slice as small as format will allow.

pressy · September 1, 2003, 6:35pm

to get 2 100 GB you have to stripe them (concart) [but i would prefer raid5.....]
what ever.....
prepair the disks:
slice 0 35,xxGB
slice 1 25mb

on each disk should be an replica (in my way s1),
metadb -a -f c0t0d0s1 cxtxdxs1 .....

now you can set up your stripe....

question:
c0t0d0 ? /pci@1f,4000/scsi@3/sd@0,0
ohh!?! isn't that your internal system (boot) disk? (or is it an ide maschine).... you should not include your system disks into the stripe... if one of the disks dies the system will die....
[[
why do you need so much space... make your system redundant..
c0t0d0 <--systemmirror--> c1t1d0
RAID5 >> the others
(1 hotspare)
]]

finster · September 2, 2003, 12:28am

Thanks.... c0t0d0 is the boot and is not include in the RAID. Need the space for program + oracle databases (large partitions due to med. network node build + collecting 30 days data on nodes)

I have played around and managed to create database replicas, and a RAID5 metadevice.

(I just picked 4 slices from 4 drives and created the metadevice for practice to get a feel for it).

# metastat
d10: RAID
State: Okay
Interlace: 64 blocks
Size: 62364843 blocks
Original device:
Size: 62365632 blocks
Device Start Block Dbase State Hot Spare
c0t1d0s6 3539 Yes Okay
c1t1d0s6 3539 Yes Okay
c1t2d0s6 3539 Yes Okay
c1t3d0s6 3539 Yes Okay

VFSTAB FILE:
# more vfstab
#device device mount FS fsck mount mount
#to mount to fsck point type pass at boot options
#
#/dev/dsk/c1d0s2 /dev/rdsk/c1d0s2 /usr ufs 1 yes -
fd - /dev/fd fd - no -
/proc - /proc proc - no -
/dev/dsk/c0t0d0s1 - - swap - no -
/dev/dsk/c0t0d0s0 /dev/rdsk/c0t0d0s0 / ufs 1 no
-
/dev/dsk/c0t0d0s6 /dev/rdsk/c0t0d0s6 /usr ufs 1 no
-
/dev/dsk/c0t0d0s3 /dev/rdsk/c0t0d0s3 /var ufs 1 no
-
/dev/dsk/c0t0d0s7 /dev/rdsk/c0t0d0s7 /export/home ufs 2
yes -
/dev/dsk/c0t0d0s5 /dev/rdsk/c0t0d0s5 /opt ufs 2 yes
-
swap - /tmp tmpfs - yes -
/dev/md/dsk/d10 /dev/md/rdsk/d10 /var/mail ufs 2 yes
-

_____________________________

Last line shows "d10" metadevice. I wanted to combine these partitions and create 1 partition (/db01).

How do I call this partition db01 and mount? I tried "/" and "/db01" in the vfstab file for mount point and it did not show up.

I have it mounting /var/mail (from someone else's example and it worked).
# df -k
Filesystem kbytes used avail capacity Mounted on
/dev/dsk/c0t0d0s0 4032504 63092 3929087 2% /
/dev/dsk/c0t0d0s6 4032504 800571 3191608 21% /usr
/proc 0 0 0 0% /proc
fd 0 0 0 0% /dev/fd
mnttab 0 0 0 0% /etc/mnttab
/dev/dsk/c0t0d0s3 4032504 114020 3878159 3% /var
swap 15718944 16 15718928 1% /var/run
swap 15719560 632 15718928 1% /tmp
/dev/dsk/c0t0d0s5 5773212 2133 5713347 1% /opt
/dev/dsk/c0t0d0s7 5040814 385457 4604949 8% /export/home
/dev/md/dsk/d10 30696394 9 30389422 1% /var/mail

__________________

Once I figure out how to do one metadevice (db01), I plan to create (db02,03, 04).

Also how can I access the new /db01 to install files?

Currently mounted to /var/mail/ and mail folder properties report 512 kb? But df -k reports full amount?

Please help again.

pressy · September 2, 2003, 3:38am

? do you already have the directory ?
mkdir /db01; mount -F ufs /dev/md/dsk/d10 /db01
you also have to create a new filesystem on the metadevice first..
newfs /dev/md/rdsk/d10
or would you like to create logical partitions?

greetings Pre�y

finster · September 3, 2003, 12:06pm

I created new fs but forgot to mkdir db01!!!!

Thanks. Worked great.

Thanks so much for the help. Partitions set up....

Could not have done it without u guys.....thanks again.

finster · September 7, 2003, 1:40am

For those following this post......in an effort to give something back......this was my outcome.....I hope it helps for future searches and questions.

Thanks again to all.

___________________________________

Storedge Multipack Setup:

Need the following minimum partitions using 6 disk array:

            db01 8GB
db02 16 GB
db03 90 GB
db04 70 GB

First disk in array = c1t1d0
Last disk in array = c1t6d0

Disk 1-6 Slices partitioned \(Concatenation\) in the following manner:

S0  Database replica \(20Mb\) \(plenty for future\)
s1  unassigned
s2  unassigned
s3  unassigned
s4  db01 \(smallest\) /db01 \(2 GB\)
s5  db02 /db02 \(3GB\)
s6  db04 /db04 \(12.8GB\)
s7  db03 \(Largest\) /db03 \(16 GB\)

Disk 1 on array will be manually partitioned using the format command and set to the above partitions.
The partition Table from disk1 will then be duplicated to the remaining 5 disks.

prtvtoc /dev/rdsk/c1t1d0s2 | fmthard -s - /dev/rdsk/c1t2d0s2
\(this will copy table from Disk1 to Disk2\)

Or write vtoc to file:

prtvtoc -h /dev/rdsk/cxtxd0s2 &gt; /export/home/vtoc_disk1

Edit the file: Delete last 2 columns \(delete only 1 column if directory not listed\)

fmthard -s /export/home/vtoc_disk1 /dev/rdsk/cxtxd0s2

Create Database Replica's: ( 1 on each disk)
metadb -a -f c1t1d0s0 c1t2d0s0 c1t3d0s0 c1t4d0s0 c1t5d0s0 c1t6d0s0

Check Replica's:

metadb -i

Create Metadevices (d10,d20,d30,d40):

metainit d10 6 1 c1t1d0s4 1 c1t2d0s4 1 c1t3d0s4 1 c1t4d0s4 1 c1t5d0s4 1 c1t6d0s4
metainit d20 6 1 c1t1d0s5 1 c1t2d0s5 1 c1t3d0s5 1 c1t4d0s5 1 c1t5d0s5 1 c1t6d0s5
metainit d40 6 1 c1t1d0s6 1 c1t2d0s6 1 c1t3d0s6 1 c1t4d0s6 1 c1t5d0s6 1 c1t6d0s6
metainit d30 6 1 c1t1d0s7 1 c1t2d0s7 1 c1t3d0s7 1 c1t4d0s7 1 c1t5d0s7 1 c1t6d0s7

Creating a File System on the Concatenation:

newfs /dev/md/rdsk/d10
newfs /dev/md/rdsk/d20
newfs /dev/md/rdsk/d30
newfs /dev/md/rdsk/d40

fsck /dev/md/rdsk/d10
fsck /dev/md/rdsk/d20
fsck /dev/md/rdsk/d30
fsck /dev/md/rdsk/d40

Check Metadevices:

metastat

Make directories in /:

mkdir db01 db02 db03 db04

Add mounts to vfstab file:

/dev/md/dsk/d10 /dev/md/rdsk/d10        /db01   ufs     2       yes     \-
/dev/md/dsk/d20 /dev/md/rdsk/d20        /db02   ufs     2       yes     \-
/dev/md/dsk/d30 /dev/md/rdsk/d30        /db03   ufs     2       yes     \-
/dev/md/dsk/d40 /dev/md/rdsk/d40        /db04   ufs     2       yes     \-

Mount Logical Partitions:

shutdown -i6 -g0 -y

or 

mount /dev/md/dsk/d10 /db01
mount /dev/md/dsk/d20 /db02
mount /dev/md/dsk/d30 /db03
mount /dev/md/dsk/d40 /db04

Check mounts and partitions:

df -k

______________________________________________
______________________________________________

*****Only If Needed

To delete metadevices: Unmount partitions:

\# umount /dev/md/dsk/d40
\# umount /dev/md/dsk/d30
\# umount /dev/md/dsk/d20
\# umount /dev/md/dsk/d10

**Remove entries rom /ect/vfstab file.

Delete metadevices:

\# metaclear d10 d20 d30 d40
d10: Concat/Stripe is cleared
d20: Concat/Stripe is cleared
d30: Concat/Stripe is cleared
d40: Concat/Stripe is cleared

Check deletion:
metastat

Delete Database Replicas:

metadb -f -d c1t1d0s0 c1t2d0s0 c1t3d0s0 c1t4d0s0 c1t5d0s0 c1t6d0s0
\(-f force, -d delete\)

Check deletion:
metadb -i