qa/distros: reinstall nvme-cli on centos 9 nodes

To work around a potential linking issue between
nvme-cli ad libnvme that prevents nvme-cli from
correctly generating a hostnqn, causing

nvme_fabrics: found same hostid edb4e426-766f-44c6-b127-da2a5b7446ef but different hostnqn hostnqn

messages in dmesg and the inability to setup nvme
loop devices

Fixes: https://tracker.ceph.com/issues/67684

Signed-off-by: Adam King <[email protected]>

Author: adk3798

qa/distros/container-hosts/centos_9.stream.yaml

@@ -9,4 +9,7 @@ overrides:
 tasks:
 - pexec:
     all:
+    # in order to work around a possible nvme-cli <-> libnvme linking issue
+    # See https://tracker.ceph.com/issues/67684
+    - sudo dnf remove nvme-cli -y
     - sudo dnf install nvmetcli nvme-cli -y

qa/distros/container-hosts/centos_9.stream_runc.yaml

@@ -8,6 +8,9 @@ overrides:
 tasks:
 - pexec:
     all:
+    # in order to work around a possible nvme-cli <-> libnvme linking issue
+    # See https://tracker.ceph.com/issues/67684
+    - sudo dnf remove nvme-cli -y
     - sudo dnf install runc nvmetcli nvme-cli -y
     - sudo sed -i 's/^#runtime = "crun"/runtime = "runc"/g' /usr/share/containers/containers.conf
     - sudo sed -i 's/runtime = "crun"/#runtime = "crun"/g' /usr/share/containers/containers.conf