test(robot): add io error test cases

[chriscchien]

Which issue(s) this PR fixes:

Add test cases

• Replica Fails When Filesystem Disk Has I/O Errors
• Replica Fails When V2 Volume Replica Has I/O Errors

test result:

• v1 https://10.115.5.5/job/private/job/longhorn-e2e-test/260/
• v2 https://ci.longhorn.io/job/private/job/longhorn-e2e-test/7350/

Issue longhorn/longhorn#12288

What this PR does / why we need it:

Special notes for your reviewer:

Additional documentation or context

[yangchiu]

Revisit this after the release.

[yangchiu]

Revisit this after v1.12.0 release.

[mergify]

This pull request is now in conflict. Could you fix it @chriscchien? 🙏

[yangchiu]

More investigation is needed. I'll need some time to clarify the questions raised in the discussion.

[chriscchien]

For v1 volume, the device map path is like below in the test code(in the QA meeting, the device mapper is mount directly from /dev/xvdh, however below method sould also equivalent for this test because since test not guarantee the extra block disk is always exist, using a v1 volume as the backing disk source is a workaround)

/dev/mapper/${dm_device_name}
  -> linear
  -> /dev/longhorn/${disk_name}  (v1 Longhorn base volume)
  -> ext4
  -> mounted at ${mount_path}
  -> added to Longhorn as filesystem disk ${disk_name}

For v2 volume, the path should be like below, however there is an issue longhorn/longhorn#13354 need to clarify first

/dev/mapper/${dm_device_name}
  -> linear
  -> /dev/xvdh   (actual disk)