Uploaded image for project: 'Couchbase Server'
  1. Couchbase Server
  2. MB-6305

default bucket ends up in limbo between delete and create

    Details

    • Type: Bug
    • Status: Resolved
    • Priority: Major
    • Resolution: Fixed
    • Affects Version/s: None
    • Fix Version/s: None
    • Component/s: None
    • Security Level: Public
    • Environment:
      CentOS 5.8 x86_64 build 705c

      hardware is a quad core, hyper-threaded CPU Intel i7. Three VMs total on the HW with only two in active use. During the time of this test, only one in active use.

      Two physical disks.

      Description

      When running automated tests that delete/create the default bucket, I'm regularly able to get it into a state where it can't be recreated saying "Bucket with given name still exists". It can't be deleted either, as a DELETE request will fail. It doesn't show in the UI or config either.

      Diag logs attached.

      1. ns-diag-20120819001246.txt.gz
        821 kB
        Matt Ingenthron
      2. ns-diag-20120819003859.txt.gz
        455 kB
        Matt Ingenthron
      No reviews matched the request. Check your Options in the drop-down menu of this sections header.

        Activity

        Hide
        ingenthr Matt Ingenthron added a comment -

        I'm still missing the technical description. What does "wait until all vbuckets are marked as ready"? Does it mean query the REST interface until some attribute hits some value? Does it mean send a stats request to all nodes looking for some value? Does it mean send some binary command and look for some specific response?

        Thanks!

        FWIW, I was incorrect in my earlier statement about SSDs. I repro'd this very easily on a Mac OS build with an SSD. Difference is mine was just doing a delete (sleep a few seconds) then a create (sleep a few seconds). It was able to get the system wedged to where I couldn't delete it, but couldn't create it.

        Show
        ingenthr Matt Ingenthron added a comment - I'm still missing the technical description. What does "wait until all vbuckets are marked as ready"? Does it mean query the REST interface until some attribute hits some value? Does it mean send a stats request to all nodes looking for some value? Does it mean send some binary command and look for some specific response? Thanks! FWIW, I was incorrect in my earlier statement about SSDs. I repro'd this very easily on a Mac OS build with an SSD. Difference is mine was just doing a delete (sleep a few seconds) then a create (sleep a few seconds). It was able to get the system wedged to where I couldn't delete it, but couldn't create it.
        Hide
        farshid Farshid Ghods (Inactive) added a comment -

        Matt,

        Im not sure if you are addressing me in the Qs or is this something ep-engine folks can share with QE and sdk team which is what i asked in my first comment

        about reproducing or not reproducing yes that is a task i will look into and assign to QE.

        Show
        farshid Farshid Ghods (Inactive) added a comment - Matt, Im not sure if you are addressing me in the Qs or is this something ep-engine folks can share with QE and sdk team which is what i asked in my first comment about reproducing or not reproducing yes that is a task i will look into and assign to QE.
        Hide
        ingenthr Matt Ingenthron added a comment -

        @farshid, I was addressing you since you seem to have a method of doing it, but we're not sure if it's the right method

        In any event, I'll hang back and wait on repro. Let me know if you need anything on that.

        RESTful flush will fix my über-issue.

        Show
        ingenthr Matt Ingenthron added a comment - @farshid, I was addressing you since you seem to have a method of doing it, but we're not sure if it's the right method In any event, I'll hang back and wait on repro. Let me know if you need anything on that. RESTful flush will fix my über-issue.
        Hide
        farshid Farshid Ghods (Inactive) added a comment -

        and i figured this is marked as closed so no one else is actually looking it .

        Show
        farshid Farshid Ghods (Inactive) added a comment - and i figured this is marked as closed so no one else is actually looking it .
        Hide
        ingenthr Matt Ingenthron added a comment -

        ah, I'd not noticed. should we reopen?

        Show
        ingenthr Matt Ingenthron added a comment - ah, I'd not noticed. should we reopen?

          People

          • Assignee:
            chiyoung Chiyoung Seo
            Reporter:
            ingenthr Matt Ingenthron
          • Votes:
            0 Vote for this issue
            Watchers:
            0 Start watching this issue

            Dates

            • Created:
              Updated:
              Resolved:

              Gerrit Reviews

              There are no open Gerrit changes