I notice that when I initially have the raid0+1 volumes registered with dm_dso_reg_tool, and then simulate a disk failure by pulling a member disk during an i/o operation, then reconnecting the disk, and then rebooting, I see that the output of dmsetup status shows that all of the member disks are active for each raid volume.
When I check out the output of dm_dso_reg_tool –m to show the list of monitored devices, it shows that the raid volumes are not registered with dmeventd. That’s fine, so I try registering all of the raid volumes, but only two out of the three volumes only ever get registered with dmeventd. Whenever I register the third volume, all three volumes become listed as not currently registered with dmeventd.
What step am I missing?
Thanks for clarifying,