coordination #105437: [Epic] Refine our testing of multipath
Create new test module for testing multipath
The FCP topology on our z/VM testing infrastructure has recently been updated and the zfcp testsuite needs a new test module to validate multipathing.
The idea is to check the status of the known infrastructure, and the status of multipathing on top of it.
AC1: Verify that there are two host bus adapters attached, and that the corresponding channels (
0.0.fa00, 0.0.fc00) are listed
# ls -l /sys/class/fc_host total 0 lrwxrwxrwx 1 root root 0 Dec 19 17:53 host0 -> ../../devices/css0/0.0.0005/0.0.fa00/host0/fc_host/host0 lrwxrwxrwx 1 root root 0 Dec 19 17:53 host1 -> ../../devices/css0/0.0.0006/0.0.fc00/host1/fc_host/host1
AC2: Based on the output of lsscsi:
# lsscsi -xxgst [0:0:0:0x4001403200000000] disk fc:0x500507630703d3b30x760f00 /dev/sda /dev/sg0 214GB [0:0:0:0x4001405000000000] disk fc:0x500507630703d3b30x760f00 /dev/sdb /dev/sg1 42.9GB [0:0:1:0x4001403200000000] disk fc:0x500507630708d3b30x761000 /dev/sdc /dev/sg2 214GB [0:0:1:0x4001405000000000] disk fc:0x500507630708d3b30x761000 /dev/sdd /dev/sg3 42.9GB [1:0:0:0x4001403200000000] disk fc:0x500507630718d3b30x771000 /dev/sde /dev/sg4 214GB [1:0:0:0x4001405000000000] disk fc:0x500507630718d3b30x771000 /dev/sdf /dev/sg5 42.9GB [1:0:1:0x4001403200000000] disk fc:0x500507630713d3b30x770f00 /dev/sdg /dev/sg6 214GB [1:0:1:0x4001405000000000] disk fc:0x500507630713d3b30x770f00 /dev/sdh /dev/sg7 42.9GB
- verify that there are 8 SCSI block devices listed (as a result of having two adapters, two hard disks and two paths to each disk, 23 = 8)
- verify that LUN
0x4001403200000000corresponds to the 214GB disk, and LUN
0x4001405000000000to the 42.9GB disk
AC3: Based on the output of multipath -l:
# multipath -l 36005076307ffd3b30000000000000132 dm-0 IBM,2107900 size=200G features='1 queue_if_no_path' hwhandler='1 alua' wp=rw `-+- policy='service-time 0' prio=0 status=enabled |- 0:0:0:1077035009 sda 8:0 active undef running |- 0:0:1:1077035009 sdc 8:32 active undef running |- 1:0:0:1077035009 sde 8:64 active undef running `- 1:0:1:1077035009 sdg 8:96 active undef running 36005076307ffd3b30000000000000150 dm-41 IBM,2107900 size=40G features='1 queue_if_no_path' hwhandler='1 alua' wp=rw `-+- policy='service-time 0' prio=0 status=enabled |- 0:0:0:1079001089 sdb 8:16 active undef running |- 0:0:1:1079001089 sdd 8:48 active undef running |- 1:0:1:1079001089 sdh 8:112 active undef running `- 1:0:0:1079001089 sdf 8:80 active undef running
AC4: Enable the validation in YaST group and Maintenance.
- verify that the multipath device with WWID
36005076307ffd3b30000000000000132corresponds to the 200G disk and the multipath device with WWID
36005076307ffd3b30000000000000150to the 40G disk
- verify that there are 4 block devices listed for each virtual multipath device, and those corresponds to the right LUN (1077035009 and 1079001089)
It might help to check this confluence article for some more insight on the output of the above commands.
This ticket requires that poo#105440 has been completed.