XCP-ng
    • Categories
    • Recent
    • Tags
    • Popular
    • Users
    • Groups
    • Register
    • Login

    bug about provoked BACKUP FELL BACK TO A FULL due to DR job

    Scheduled Pinned Locked Moved Backup
    4 Posts 3 Posters 56 Views 3 Watching
    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.
    • P Online
      Pilow
      last edited by Pilow

      Hi,
      latest XOA 6.0.3

      I found a case where I can provoke the 8ae00fdb-ab5d-4edc-ba05-95e252a2a5dd-{B9996169-F642-4FA6-883B-A1F757E52FB9}.png
      I do not understand why it happens, but it is happening anyway.

      We have a normal DELTA backup job with 6 VMs in it. say VM1 to VM6.
      This job has advanced options of NBD+CBT checked. NBD network exists to do the backup.
      Normal snapshot mode. Merge Synchronously in advanced is checked.

      Another job : CONTINUOUS REPLICATION, same 6 VMs, same advanced options checked

      and a last job : DISASTER RECOVERY of a subset of two VMs : VM4 and VM6
      ZSTD compression, normal snapshot mode, merge backup synchronously. 1 point retention (a full is done each time, as intended)

      All 3 jobs are in sequence, they do not overlap. first backup, second replica, third DR.

      All 3 jobs execute with success, but in the BACKUP and REPLICA jobs :
      only VM4 and VM6 have 703c771e-d223-45ac-aff7-6487238be1b0-{B9996169-F642-4FA6-883B-A1F757E52FB9}.png with "delta" backup type at the end.

      the fact that the DR job is done one these two VMs provoke this behavior. this is not wanted.

      What we see in REPLICA job log :
      7d8bdcda-ce93-4541-8513-f83ea28ab46e-{929E2473-40BB-4AA3-AB8D-BDCFF75B3744}.png

      And in backup JOB log :
      47ce2123-3492-45fd-a61a-6ab5307b68d9-{9C9CD199-AF57-4B41-9C87-21762C5C2CDB}.png

      I checked on VM VDI after the DR job is finished and CBT is still checked
      29f400f1-dcd2-43cc-b235-9de20f072ed5-{FBF26978-1EFE-4610-87B7-C06F69C7EEA0}.png

      Why does a DR job resets something that make other two jobs fall back to full on these VDIs ?

      Bastien NolletB 1 Reply Last reply Reply Quote 0
      • Bastien NolletB Offline
        Bastien Nollet Vates 🪐 XO Team @Pilow
        last edited by

        Thanks @Pilow for the report, I'll try to reproduce on my side to get a better undertanding of what's creates the fallback.

        P 1 Reply Last reply Reply Quote 0
        • P Online
          Pilow @Bastien Nollet
          last edited by Pilow

          @Bastien-Nollet nice.

          I can confirm with the dates of the backup logs that it is since the exact date of beginning of DR job that this problem appears on the two other jobs.

          when not doing NBD+CBT, the backup & replica jobs keep the "full" snapshot on the VM.
          DR do not do that, it delete the snapshot.

          I guess this is a silent failure of the way DR manages its snapshot deletion that could delete the CBT bitmap ? and then, baaam fall back to full on other jobs relying on said bitmap... ?

          this is a run of the DR job
          ed0a138a-bf0a-40d1-b073-5542f9e79d02-image.png

          no notion of NBD/CBT in DR job advanced parameters, it treats that differently
          6c5e772d-eebb-4918-8f49-4a36cafcf7df-image.png

          florentF 1 Reply Last reply Reply Quote 0
          • florentF Offline
            florent Vates 🪐 XO Team @Pilow
            last edited by

            I am summoning @storage (especially G. R. )

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