Uploaded image for project: 'OpenShift Virtualization'
  1. OpenShift Virtualization
  2. CNV-55705

qemu-pr-helper is not working with new SCSI devices and requires a restart of the virt-handler pod

XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Unresolved
    • Icon: Undefined Undefined
    • CNV v4.17.32
    • CNV v4.17.3
    • Storage Ecosystem
    • None
    • Quality / Stability / Reliability
    • 0.42
    • False
    • Hide

      None

      Show
      None
    • False
    • None
    • CNV Storage 267
    • Important
    • None

      Description of problem:

      Mapped a new disk from an iSCSI target to the node:

      # iscsiadm -m session -P 3 |grep sd
       Attached scsi disk sdb State: running
      
      # [root@10 ~]# lsblk  |grep sdb
      sdb        8:16   0    30G  0 disk
      ├─sdb1     8:17   0    16M  0 part
      └─sdb2     8:18   0     1K  0 part

      Created a PV on this device, and then it was attached to the VM with the reservation set to 'true.' Tried running a validation test for the Windows Failover Cluster, but the validation failed. The strace of qemu-pr-helper reported that it was unable to see sdb.

      1455447 08:47:00.142969 read(14</sys/devices/platform/host13/session17/target13:0:0/13:0:0:0/state>, "running\n", 19) = 8
      1455447 08:47:00.143010 close(14</sys/devices/platform/host13/session17/target13:0:0/13:0:0:0/state>) = 0
      1455447 08:47:00.143048 openat(AT_FDCWD</>, "/dev/sdb", O_RDONLY) = -1 ENOENT (No such file or directory)
      1455447 08:47:00.143107 close(12</dev/disk-chocolate-crane-84>) = 0

      "pr-helper" container is unable to see sdb:

      [core@10 ~]$ oc rsh -c pr-helper virt-handler-j5lff
      sh-5.1# ls /dev/sd*
      /dev/sda  /dev/sda1  /dev/sda2    /dev/sda3  /dev/sda4

      If I restart the virt-handler pod, and stop and start the VMs (required because of CNV-55559), the validation is successful. So a restart of the pod is required for virt-handler to see the new devices.

      Version-Release number of selected component (if applicable):

      OpenShift Virtualization 4.17.3

      How reproducible:

      100%

      Steps to Reproduce:

      1. Map new device on the node. Ensure that virt-handler is not restarted after mapping the new device.
      2. The new device will not visible in the pr-helper container.
      3. Attach the disk to the VM with "reservation: true". 
      4. Try doing a WSFC validating test from  Windows or run sg_persist from a RHEL VM. It will fail.

      Actual results:

      qemu-pr-helper is not working with new SCSI devices and requires a restart of virt-handler pod

      Expected results:

       

      Additional info:

       

              afrosirh Alice Frosi
              rhn-support-nashok Nijin Ashok
              Kevin Alon Goldblatt Kevin Alon Goldblatt
              Votes:
              0 Vote for this issue
              Watchers:
              12 Start watching this issue

                Created:
                Updated: