Example 1: Logical Device With SVM
To illustrate what has been presented, this section describes a method to access a replicated set of data, created using an enterprise class subsystem: the StorEdge SE99x0 systems. ShadowImage is used to replicate the LUNs. The data to access is configured as a metadevice using SVM patched at the latest level. The primary metadevice is called d101 and is a soft partition created on top of metadevice d100: RAID-0 of four SE99x0 LUNs. All of these metadevices are part of the metaset labset.
In this example, the primary and secondary volumes (metadevices) are accessed from two different hosts (the primary host is storage10 and the secondary host is storage26). In this situation, the primary host has access to the primary LUNs only, and secondary host sees only the secondary LUNs. This constraint forces you to reconstruct the metaset and metadevices on the secondary site before accessing the data. There is no possibility of importing or exporting the metaset from one host to the other (take and release ownership of a metaset implies that every disk is visible on both hosts).
d100, RAID 0 of 4 LUNs
d101, softpartition on top of d100
LUNS visible from storage103
LUNS visible from storage26
ShadowImage is set up so that LUNs are paired, under the consistency group LAB, as follows:
Primary Disk -->
As described in the previous section, to access the replicated data, you must insure that every layer of the I/O stack is correctly set up. In this example, the steps would be divided as follows:.
Physical layerensure the replicated disks are consistent and accessible.
Driver layerdetect the replicated disks.
LVM Layerreconstruct the replicated metasets and metadevices.
File system layermake the replicated file system consistent.
Application layermake the data ready for the application.
To Ensure Consistent and Accessible Replicated Disks in the Physical Layer
Suspend the replication:
Before accessing the replicated LUNs, stop (suspend) the replication and make sure every LUN is in a suspended (psus) state:
root@storage103 # pairsplit -g LAB root@storage103 # pairevtwait -g LAB -s psus -t 1800
Of course, pairspli must be issued when all the LUNs are already synchronized (state PAIR). Failing to do so will result in corrupted secondary devices
To Detect the Replicated Disks in the Driver Layer
Scan the disks and verify that they are all visible and accessible from the secondary host.
This is achieved using the Solaris OS command devfsadm, which scans I/O buses for new devices and reads the partition table of each disk:
root@storage26 # devfsadm
To Reconstruct the Replicated Metasets and Metadevices in the LVM Layer
Modify the primary metaset configuration to reflect the new devices, and apply the modified configuration to a newly created metaset.
Create a metaset on secondary host:
Populate the new metaset with cloned disks:
Create new configuration for the secondary metaset.
On the secondary host, create a metadevice configuration file called /etc/lvm/md.tab containing the previous output with the correct secondary LUNs.
Apply the metadevice configuration file to the replicated host:
root@storage26 # metaset -s labset -a -h storage26
root@storage26 # metaset -s labset -a \ c3t500060E8000000000000ED160000020Ad0\ c3t500060E8000000000000ED160000020Bd0 \ c3t500060E8000000000000ED160000020Cd0 \ c3t500060E8000000000000ED160000020Dd0 \
Start by obtaining the metadevice configuration of the primary host:
root@storage103 # metaset -s labset -p labset/d101 -p labset/d100 -o 1 -b 10485760 labset/d100 1 4 c8t500060E8000000000000ED1600000200d0s0 \ c8t500060E8000000000000ED1600000201d0s0 \ c8t500060E8000000000000ED1600000202d0s0 \ c8t500060E8000000000000ED1600000203d0s0 -i 32b
The order of appearance must be respected:
root@storage26 # cat /etc/lvm/md.tab labset/d101 -p labset/d100 -o 1 -b 10485760 labset/d100 1 4 c3t500060E8000000000000ED160000020Ad0s0 \ c3t500060E8000000000000ED160000020Bd0s0 \ c3t500060E8000000000000ED160000020Cd0s0 \ c3t500060E8000000000000ED160000020Dd0s0 -i 32b
root@storage26 # metainit -s labset -a labset/d100: Concat/Stripe is setup labset/d101: Soft Partition is setup root@storage26 #
To Make the Replicated File System Consistent in the File System Layer
Check the consistency of the file system:
Mount the file system:
root@storage26 # fsck /dev/md/labset/rdsk/d101
The fsck lists the corrupted files. Action must be taken to recover them. This operation makes sense in case of a crash. During a crash, some files might be corrupted (files in creation and modification mode).
root@storage26 # mount /dev/md/labset/dsk/d101 /mnt/LAB
To Make the Data Ready for the Application in the Application Layer
At this stage, you can consider the replicated data accessible. Some application specific actions might take place, such as modifying configuration files, links, or other clean up and recover processes.