none
Cluster Shared Volumes switch to Redirected Access when adding new node

    Question

  • Hello,

    Here is a little background: I built a 3 nodes Hyper-V 2008 R2 cluster, first checking that all seven of my volumes (stored on a direct attached Dell MD3200+1200) were visible to my nodes and initialized (GPT + NTFS) then creating the cluster. Everything was working fine, the cluster validated, cluster shared volumes were happily Online and I could do live migrations between the nodes and do P2V migrations.

    I then set up a VM to host our DPM 2010 and wanted to do bare metal backups of my Hyper-V nodes. I read I needed to install two hotfixes first, which I did on Node1 as a test. After restart the CVS switched to Redirected Access. I figured it was because of the hotfixes, so I uninstalled both of them, restarted, but still got the Redirected Acess problem. I removed Node1 from the cluster and immediately CVS were back Online and everything was fine for Node2 and Node3. Now when trying to add a new node I get a Redirected Access on all volumes after a few seconds. Removing the new node immediately corrects the problem. New nodes see the disks as uninitialized and can't access them which makes sense before they're added to the cluster but they should be able to see them after that...

    I have restarted both working nodes in turn, deactivated FsDepends that was running only on Node3, assigned volumes to Node2 then to Node3 (I can't assign them to new nodes of course), taken offline individual CSV, nothing seems to be working.

    Does anyone have an idea about what is happening and what I could do to fix the problem?


    • Edited by peterclo Friday, November 18, 2011 1:24 PM
    Friday, November 18, 2011 1:24 PM

Answers

  • I just found out what happened. Here are the steps:

    1. I installed the three physical servers in a row using automated steps with MDT and automatic updates from WSUS (both located on the same server)

    2. Things were working nicely, I did P2V migrations of a web server and of my admin server (the one with MDT+WSUS).

    3. I installed a few VMs, booting from the network on MDT. Everything looked pretty normal.

    4. I installed DPM on a VM, something broke because of the hotfixes. I ended up having to remove CasNode1 from the cluster.

    5. I reinstalled CasNode1, booting from the nework on MDT.

    6. Still getting Redirected Access everywhere. I read tens of pages and tips (including the one you linked Samir!), trying things and generally failing.

    7. I had an epiphany and checked my WSUS server: it kept crashing on startup!

    8. I stopped my admin VM and booted up the old physical server. WSUS working fine here. I checked update on the new node: it found 44 of them! Updated, rebooted: everything is Online now... I'm still not clear why WSUS keeps crashing on the VM but I wish I had found out it did before spending so much time trying to find a fix :)

    Friday, November 18, 2011 4:41 PM

All replies

  • Who is  the Owner of CSV disks, after and befor adding the new node ? or Owners ?

    Can you post which CSVs are redirected?

    A screen shot is prefferred ?

     


    Regards, Samir Farhat Infrastructure Consultant
    Friday, November 18, 2011 1:43 PM
  • Hi Samir, thanks for your answer.

    Here is a "before" screenshots of the CSVs:

    I then tried adding a new node, which worked fine (even the Clustering all shared disks step) but then:

    Evicting the new node brings all CSVs status back to Online.

    I've tried assigning all CSVs to CasNode3 and then adding a new node, with the same results.

    Shutting down CasNode2 or 3 while the new node is up keeps all CSVs in Redirected Access.

    Friday, November 18, 2011 2:02 PM
  • Try to look the following link, i used it to troubleshot a CSV redirected access mode

    http://blogs.technet.com/b/askcore/archive/2010/12/16/troubleshooting-redirected-access-on-a-cluster-shared-volume-csv.aspx

    Are Client for Microsoft Networks and File and Printer Sharing for Microsoft Networks protocols enabled in your network cards?


    Regards, Samir Farhat Infrastructure Consultant
    Friday, November 18, 2011 3:51 PM
  • I just found out what happened. Here are the steps:

    1. I installed the three physical servers in a row using automated steps with MDT and automatic updates from WSUS (both located on the same server)

    2. Things were working nicely, I did P2V migrations of a web server and of my admin server (the one with MDT+WSUS).

    3. I installed a few VMs, booting from the network on MDT. Everything looked pretty normal.

    4. I installed DPM on a VM, something broke because of the hotfixes. I ended up having to remove CasNode1 from the cluster.

    5. I reinstalled CasNode1, booting from the nework on MDT.

    6. Still getting Redirected Access everywhere. I read tens of pages and tips (including the one you linked Samir!), trying things and generally failing.

    7. I had an epiphany and checked my WSUS server: it kept crashing on startup!

    8. I stopped my admin VM and booted up the old physical server. WSUS working fine here. I checked update on the new node: it found 44 of them! Updated, rebooted: everything is Online now... I'm still not clear why WSUS keeps crashing on the VM but I wish I had found out it did before spending so much time trying to find a fix :)

    Friday, November 18, 2011 4:41 PM
  • Nice that you fixed it. But we usually prefer theses information when posting the thread. More info, more easy to troubleshoot.
    Regards, Samir Farhat Infrastructure Consultant
    Friday, November 18, 2011 5:04 PM