projects / thirdparty / mdadm.git / blame
| Commit | Line | Data |
|---|---|---|
| beb71de0 ZL |
1 | # Prerequisite: |
| 2 | # 1. The clustermd_tests/ cases only support to test 2-node-cluster, cluster | |
| 3 | # requires packages: 'pacemaker+corosync+sbd+crmsh', all packages link at | |
| 4 | # "https://github.com/ClusterLabs/", and also requires dlm resource running | |
| 5 | # on each node of cluster. | |
| 6 | # For quick start HA-cluster with SUSE distributions, refer to the chapter 6-8: | |
| 7 | # https://www.suse.com/documentation/sle-ha-12/install-quick/data/install-quick.html | |
| 8 | # For Redhat distributions, please refer to: | |
| 9 | # https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/7/html-single/high_availability_add-on_administration/index | |
| 10 | # 2. Setup ssh-access with no-authorized mode, it should be: | |
| 11 | # # 'ssh $node1 -l root ls' and 'ssh $node2 -l root ls' success on any node. | |
| 12 | # 3. Fill-up node-ip part and disks part as following. | |
| 13 | ||
| 14 | # Set node1 as the master node, the cluster-md cases should run on this node, | |
| 15 | # and node2 is the slave node. | |
| 16 | # For example: | |
| 17 | # NODE1=192.168.1.100 (testing run here) | |
| 18 | # NODE2=192.168.1.101 | |
| 19 | NODE1= | |
| 20 | NODE2= | |
| 21 | ||
| 22 | # Provide the devlist for clustermd-testing, alternative: if set the step 1, | |
| 23 | # don't set step 2, and vice versa. | |
| 24 | # 1. Use ISCSI service to provide shared storage, then login ISCSI target via | |
| 25 | # to ISCSI_TARGET_ID and ISCSI_TARGET_IP on iscsi clients, commands like: | |
| 26 | # Execute on iscsi clients: | |
| 27 | # 1) discover the iscsi server. | |
| 28 | # # iscsiadm -m discovery -t st -p $ISCSI_TARGET_IP | |
| 29 | # 2) login and establish connection. | |
| 30 | # # iscsiadm -m node -T $ISCSI_TARGET_ID -p $ISCSI_TARGET_IP -l | |
| 31 | # Note: | |
| 32 | # On ISCSI server, must create all iscsi-luns in one target_id, recommend more | |
| 33 | # than 6 luns/disks for testing, and each disk should be: 100M < disk < 800M. | |
| 34 | # 2. If all cluster-nodes mounted the same disks directly, and the devname are | |
| 35 | # the same on all nodes, then put them to 'devlist'. | |
| 36 | ||
| 37 | # For example: (Only set $ISCSI_TARGET_ID is enough if iscsi has already connected) | |
| 38 | # ISCSI_TARGET_ID=iqn.2018-01.example.com:clustermd-testing | |
| 39 | # ISCSI_TARGET_IP=192.168.1.102 | |
| 40 | ISCSI_TARGET_ID= | |
| 41 | ||
| 42 | #devlist=/dev/sda /dev/sdb /dev/sdc /dev/sdd | |
| 43 | devlist= |