Lucent Technologies CentreVu Call Management System Release 3 Version 8 Guide

							  Maintaining Mirrored Systems  CentreVu CMS Release 3 Version 8 Disk-Mirrored Systems
Adding new disks to a standard mirrored system65
13. Add the new disks to the existing metadevices by entering the 
following command:
(where, again, 
 and  are the device names 
of the disks you have added).
The system responds with the following:
14. Restart CMS by entering the cmsadm 
or cmssvc 
command and 
selecting the cms_run
 option. When the run
 option menu displays, 
select start cms
.
15. Check the available disk space by entering a df -k 
command. For 
example:
Compare the 
avail figure for /cms to the figure you recorded in 
St e p 8
; it should show an increase roughly equal to one-half of the 
total capacity of the drives added.
# olds -mirrored -setup   
Warning: Current Disk has mounted partitions.
device: c0t6d0 will not be used
. . .
d19: component is attached
d20: component is attached
...
d21: submirror d20 is attached
Success, growing d21
Success, the capacity of the /cms mirror has increased.
#
#df-k
Filesystem kbytes used avail capacity Mounted on
/dev/md/dsk/d13 772038 538683 156155 78% /
/proc 0 0 0 0% /proc
fd 0 0 0 0% /dev/fd
/dev/md/dsk/d21 3059955 192497 4967458 7% /cms 
						

							  Maintaining Mirrored Systems  CentreVu CMS Release 3 Version 8 Disk-Mirrored Systems
Adding new disks to a nonstandard mirrored system66
Adding new disks to a nonstandard mirrored system3
A “nonstandard” disk drive addition is one that cannot take advantage of 
advanced features of the 
olds setup scripts because the disk 
configuration is not standard. The procedure is as follows:
1. Turn off CMS. (From the command line, enter the cmsadm 
or 
cmssvc 
command, and select the cms_run
 option. When the run
 
option menu displays, select Turn off cms
.)
2. Halt the system by entering the following command:
3. Install the new disk drives by plugging them into the appropriate 
empty drive slots.
Follow the instructions you received with the disk drives or with your 
basic system, or see the hardware installation or maintenance and 
troubleshooting document for your platform and version of CMS.
4. Reboot the system by entering the following command:
5. Determine the total amount of disk space available to /cms by 
entering the following command:
The system responds by displaying file system information for /cms. 
For example:
Check the response. Look for the “avail” figure for 
/cms (shown as 
“yyyyyy” in the example above). Record that figure here:
/cms “avail” figure:______________________
6. Partition the new disks. Follow the instructions in “Partitioning disks 
with the format command” on page 76. Double check the partitioning 
before continuing.# /usr/sbin/shutdown -y -g0 -i0
ok boot -r
# df -k /cms
Filesystem kbytes used avail capacity Mounted on
/dev/md/dsk/d21 xxxxxxxx 9 yyyyyy 0% /cms 
						

							  Maintaining Mirrored Systems  CentreVu CMS Release 3 Version 8 Disk-Mirrored Systems
Adding new disks to a nonstandard mirrored system67
7. Add the 
olds directories to your path:
8. Create new versions of the setup files by entering the following 
command:
 and  represent the device names of the 
disk drives added (remember that in a mirrored system, disk drives 
are always added in pairs). For example:
9. Load the file 
/olds/md.tab.new into an editor and check to make 
certain it names all
 of your disks, including the ones you just added. 
(For an example of the format of an md.tab file, see “Examples of 
valid system files” on page 79.)
10. Save a copy of the existing 
md.tab file by entering the following 
command:
11. Replace the existing 
md.tab file with the one you created in Step 8:
12. Attach one of the newly installed disk drives to each of the d21 
submirrors (d19 and d20):
where 
 and  represent the device names of 
the disk drives you have just added. For example:# PATH=$PATH:/usr/opt/SUNWmd/sbin:/olds
# export PATH
# olds -mirrored -mk_files 
# olds -mirrored -mk_files c0t10d0 c0t11d0
# cp /etc/opt/SUNWmd/md.tab /etc/opt/SUNWmd/md.tab.mirror
# cp /olds/md.tab.new /etc/opt/SUNWmd/md.tab
# metattach d19
# metattach d20
# metattach d19 c0t10d0s1
# metattach d20 c0t11d0s1 
						

							  Maintaining Mirrored Systems  CentreVu CMS Release 3 Version 8 Disk-Mirrored Systems
Adding new disks to a nonstandard mirrored system68
13. Grow the 
/cms file system by entering the following command:
If the operation should fail, do the following:
a. Restore the original md.tab file by entering the following 
command:
b.Look up the error message in Chapter 4, “Troubleshooting.”
14. Enter the following command:
Examine the output to make certain the disk drives you just added 
are part of the configuration.
15. Enter the following command:
The system responds by displaying file system information for /cms. 
For example:
Check the response. Look for the “avail” figure for /cms (shown as 
“yyyyyy” in the example above). That figure should be roughly twice 
the figure recorded in Step 5.# growfs -M /cms /dev/md/rdsk/d21
# cp /etc/opt/SUNWmd/md.tab.mirror /etc/opt/SUNWmd/md.tab
# metastat
# df -k /cms
Filesystem kbytes used avail capacity Mounted on
/dev/md/dsk/d21 xxxxxxxx 9 yyyyyy 0% /cms 
						

							  Maintaining Mirrored Systems  CentreVu CMS Release 3 Version 8 Disk-Mirrored Systems
Replacing a faulty disk69
Replacing a faulty disk3
This section explains how to replace a faulty disk. The system does not 
need to be powered off or rebooted to perform this procedure.
Step 1: Identify the 
faulty disk
3
A disk problem is usually indicated in the /var/adm/messages file. The 
following lines, for example, indicate a disk problem: 
Those lines, however, do not tell us which disk has the problem. One way 
of finding out is to enter an ls -l /dev/dsk/c* 
command and 
search the output for a device description matching that in the warning 
message. For example:
That the ”/iommu@0....” information matches the same information in the 
warning message indicates that disk c0t2d0 is the faulty disk.Jun 12 16:27:08 leopard unix: WARNING: 
Jun 12 16:27:08 leopard unix: Error for command 
read(10) Error Level: R
Jun 12 16:27:09 leopard unix: retryable
Jun 12 16:27:09 leopard unix: Requested Block 0, 
Error Block: 0
Jun 12 16:27:09 leopard unix: Sense Key: Media Error
Jun 12 16:27:09 leopard unix: Vendor SEAGATE:
Jun 12 16:27:09 leopard unix: ASC = 0x31 (medium 
format corrupted), ASCQ
= 0x0, FRU = 0x9
Jun 12 16:27:09 leopard unix: WARNING: 
/sbus@3,0/SUNW,fas@3,8800000/sd@2,0
# ls -l /dev/dsk/c*
.
.
.
lrwxrwxrwx   1 root     root          50 Apr 24 15:21 
/dev/dsk/c0t2d0s0 -> ../../devices/sbus@3,0/SUNW,fas 
@3,8800000/sd@2,0:a
lrwxrwxrwx   1 root     root          50 Apr 24 15:21 
/dev/dsk/c0t2d0s1 -> ../../devices/sbus@3,0/SUNW,fas 
@3,8800000/sd@2,0:b
.
.
.
# 
						

							  Maintaining Mirrored Systems  CentreVu CMS Release 3 Version 8 Disk-Mirrored Systems
Replacing a faulty disk70
Step 2: Identify the 
submirrors
3
1. Add the Solstice DiskSuite directories to your path:
2. Enter a metastat
 command:
3. Scan the output of the metastat command to find the name of the 
faulty disk. The following lines, for example, indicate that the faulty 
c0t2d0 disk is in d20:
Be sure to check all
 the submirrors. If the disk is also listed under 
d11 or d12, it is a boot disk belonging to two different submirrors. 
Replacing a boot disk involves some special considerations; those 
considerations are noted in this procedure where they apply.
4. Record below the metadevices to which the disk belongs (for 
example, d11 and d19):
d____ d____ # export PATH=$PATH:/usr/opt/SUNWmd/sbin:/olds
# metastat
# metastat
.
.
d20: Submirror of d21
Size: 14960160 blocks
Stripe 0:
Device Start Block Dbase State Hot Spare
c0t2d0s1 0 No Okay
Stripe 1:
Device Start Block Dbase State Hot Spare
c0t4d0s3 0 No Okay         
.
. 
						

							  Maintaining Mirrored Systems  CentreVu CMS Release 3 Version 8 Disk-Mirrored Systems
Replacing a faulty disk71
Step 3: Detach the 
submirrors
3
To detach a submirror means to put it off line so you can access its 
constituent disk drives. You detach a submirror with a metadetach
 
command. The command is in the format:
metadetach
 
where  is the mirror and  is the submirror to be 
detached. 
To continue our example, since c0t2d0 is part of d20, which is a submirror 
of d21, the detach command would be as follows:
If the faulty disk is a boot disk, a metastat
 command also shows the 
disk as part of the d11 or d12 submirror. Be sure to detach both the 
submirrors of which it is a member. For example:# metadetach d21 d20
d21: submirror d20 is detached
# metadetach d13 d12
d13: submirror d12 is detached
# metadetach d21 d20
d21: submirror d20 is detached 
						

							  Maintaining Mirrored Systems  CentreVu CMS Release 3 Version 8 Disk-Mirrored Systems
Replacing a faulty disk72
Step 4: Remove 
state database 
replicas (boot disks 
only)
3
When you replace a boot disk, you must remove the state database 
replicas from the old disk and recreate them on the new one. A boot disk 
is one that is a member of metadevice d11 or d12 (see “Step 2: Identify 
the submirrors” on page 70).
If you fail to remove old replicas, the software will have an 
inaccurate picture of the replicas existing on your system and will 
attempt to write to nonexistent database replica files.Match up the 
device description lines to determine which disk drive has 
problems.
In general, the procedure is a three-step process:
1. Issue a metadb -i 
command to find out which replicas to 
remove.
2. Issue a metadb -d 
command to do the actual deletion.
3. Issue another metadb -i 
command to verify the deletion.
If the faulty disk were c0t1d0, for example, the replica deletion might 
resemble the following series of commands and responses:
Step 5: Replace the 
faulty disk drive
3
1. Remove the faulty disk drive.
2. Install the new disk drive into the same slot the faulty drive came out 
of. For installation instructions, see the documentation that came 
with the disk drive or the Lucent hardware installation manual.
3. Partition the new disk using the format
 command. See 
“Partitioning disks with the format command” on page 76
. Make 
sure you partition the new disk drive!
# metadb -i
flags first blk block count
a m p luo 16 1034 /dev/dsk/c0t0d0s1
a p luo 1050 1034 /dev/dsk/c0t0d0s1
a p luo 2084 1034 /dev/dsk/c0t0d0s1
a p luo 3118 1034 /dev/dsk/c0t0d0s1
a p luo 16 1034 /dev/dsk/c1t4d0s1
a p luo 1050 1034 /dev/dsk/c1t4d0s1
a p luo 2084 1034 /dev/dsk/c1t4d0s1
# metadb -d c1t4d0s1
# metadb -i
flags first blk block count
a m p luo 16 1034 /dev/dsk/c0t0d0s1
a p luo 1050 1034 /dev/dsk/c0t0d0s1
a p luo 2084 1034 /dev/dsk/c0t0d0s1
a p luo 3118 1034 /dev/dsk/c0t0d0s1 
						

							  Maintaining Mirrored Systems  CentreVu CMS Release 3 Version 8 Disk-Mirrored Systems
Replacing a faulty disk73
Step 6: Recreate 
the state database 
replicas
3
1. Recreate the primary boot disk replicas you removed earlier by 
entering the following command:
2. Recreate the secondary boot disk replicas you removed earlier by 
entering the following command:
3. Verify that the replicas were created:
The system lists the state database replicas currently on the system. 
The following, for example, shows three replicas on c0t4:
In the above example, there are 4 replicas listed for the primary boot 
disk 
(c0t0d0s1), and 3 replicas listed on the secondary disk 
(c1t4d0s1)
Do not try to reboot a system when there are fewer than two state 
database replicas.
#metadb -a -c4 -f /dev/rdsk/devname
#metadb -a -c3 -f /dev/rdsk/devname
# metadb -i
flags first blk block count
a m p luo 16 1034 /dev/dsk/c0t0d0s1
a p luo 1050 1034 /dev/dsk/c0t0d0s1
a p luo 2084 1034 /dev/dsk/c0t0d0s1
a p luo 3118 1034 /dev/dsk/c0t0d0s1
a p luo 16 1034 /dev/dsk/c1t4d0s1
a p luo 1050 1034 /dev/dsk/c1t4d0s1
a p luo 2084 1034 /dev/dsk/c1t4d0s1
!WARNING: 
						

							  Maintaining Mirrored Systems  CentreVu CMS Release 3 Version 8 Disk-Mirrored Systems
Replacing a faulty disk74
Step 7: Reattach 
the submirrors
3
1. Enter a metainit command to reinitialize the submirror. For example:
If you receive an “already set up” message, ignore it.
2. Reattach the submirror by issuing a metattach
 command. For 
example:
If the disk you are replacing is a boot disk, be sure to reattach 
both the submirrors of which it is a member. For example:
Step 8: Reboot the 
system (boot disks 
only)
3
You may delay this step until a more convenient time.
If the replaced disk is a boot disk, you will have to reboot the system at 
your convenience, typically during low busy hours. The reboot is 
necessary because the 
/etc/system file was modified when the state 
database replicas were recreated.# metainit d20
#
# metattach d21 d20
#
# metattach d13 d12
# metattach d21 d20