XCP-ng

    • Register
    • Login
    • Search
    • Categories
    • Recent
    • Tags
    • Popular
    • Users
    • Groups

    Auto coalescing of disks

    Xen Orchestra
    coalesce storage
    3
    14
    1252
    Loading More Posts
    • Oldest to Newest
    • Newest to Oldest
    • Most Votes
    Reply
    • Reply as topic
    Log in to reply
    This topic has been deleted. Only users with topic management privileges can see it.
    • C
      chrispage1 @r1 last edited by

      @r1 They are reporting to be at least, all green on the UI -

      85dd30d1-768c-4de5-8939-4d8d49571b44-image.png

      Any way I can check further?

      Chris.

      1 Reply Last reply Reply Quote 0
      • olivierlambert
        olivierlambert Vates 🪐 Founder & CEO 🦸 last edited by olivierlambert

        By default the UI will only display running hosts. Be sure to have an empty search field.

        C 1 Reply Last reply Reply Quote 0
        • C
          chrispage1 @olivierlambert last edited by

          @olivierlambert thanks - that's definitely all the hosts. Only three on this system. Any way to check each host at a deeper level?

          Chris

          1 Reply Last reply Reply Quote 0
          • olivierlambert
            olivierlambert Vates 🪐 Founder & CEO 🦸 last edited by

            Input/output error isn't a good sign 😕

            Try to reboot all your hosts and see if the issue persist. Also check dmesg.

            C 1 Reply Last reply Reply Quote 0
            • C
              chrispage1 @olivierlambert last edited by chrispage1

              @olivierlambert ah - would the input/output error be on the nodes itself or the SR?

              We have seen this input/output error before when attempting to snapshot or migrate VM's -

               "code": "SR_BACKEND_FAILURE_81",
                "params": [
                  "",
                  "Failed to clone VDI [opterr=['XENAPI_PLUGIN_FAILURE', 'multi', 'CommandException', 'Input/output error']]",
                  ""
                ],
              

              I can reboot the nodes but would be a little hesitant due to potential downtime, etc. Would restarting the tool stack help?

              dmesg isn't reporting anything particularly concerning.

              Appreciate your support on this,
              Chris.

              1 Reply Last reply Reply Quote 0
              • olivierlambert
                olivierlambert Vates 🪐 Founder & CEO 🦸 last edited by

                It's on the SR here. The storage code is reporting the issue, so that's why coalesce isn't happening.

                C 1 Reply Last reply Reply Quote 0
                • C
                  chrispage1 @olivierlambert last edited by

                  @olivierlambert thanks. The only other thing - I was using templates so I notice on my storage that one block consists of 6 base copies & 3 VDI's.

                  In this format, I have issues migrating and snapshotting them. However, if I power them off, snapshot, run a full clone, and use the new clone they work fine (migratable, snapshottable, bootable). Any ideas why this might be and could this be the cause of the coalesce issue?

                  Thanks,
                  Chris.

                  1 Reply Last reply Reply Quote 0
                  • C
                    chrispage1 last edited by

                    @olivierlambert Hi, I've been keeping an eye on dmesg and getting the following -

                    [Thu Apr 8 20:26:51 2021] device-mapper: ioctl: unable to remove open device VG_XenStorage--767a9d9c--55aa--0898--fe10--400ee856bdce-clone_ba7ea2dd--4684--4cc9--9ccd--ba1909cda656_1

                    I don't know if this could possibly be related to the exceptions I am seeing?

                    Thanks,
                    Chris.

                    1 Reply Last reply Reply Quote 0
                    • olivierlambert
                      olivierlambert Vates 🪐 Founder & CEO 🦸 last edited by

                      You have weird issues on your storage (I/O error on badly named VHD, hard to tell without investigating further)

                      C 1 Reply Last reply Reply Quote 0
                      • C
                        chrispage1 @olivierlambert last edited by

                        @olivierlambert thanks - I'll continue to look into it.

                        1 Reply Last reply Reply Quote 0
                        • First post
                          Last post