This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

Slave stuck in "RESERVED" after up2date to 9.506-2 - master still on 9.505-4

Hello,

 

I tried up2date a HA Cluster running 9.505-4 via WebGUI.

The slave got stuck in "up2date"-state for hours so I decided to reboot it.

When it came back it was 9.506-2 and rejoined the HA cluster but it is now in state "RESERVED" (No, I did not use that up2date-option to keep one node in reserved mode an there is no "upgrade this node"-button now, only shut down and reboot)

I can ssh to that slave node but I have no idea what to do now, rebooting it over did not help. Do I really have to destroy that cluster? If so, what are the default IPs a.s.o to get access to that node for rejoining it? An "ifconfig -a" does not show any hint on that.

 

Thank you.

 

Chris



This thread was automatically locked due to age.
Parents
  • Chris, is the Master at 9.506 now?  If so, then it seems like the process almost completed normally.

    The first step in an HA Up2Date is for the Master (let's say this is node 1) to Up2Date the Slave (node 2).  Once the node 2 is Up2dated, rebooted and synced with the Master node, node 2 becomes Master.  Finally, node 2 Up2Dates node 1, node 1 then reboots and syncs with node 2.

    I would get Sophos Support involved to confirm that you don't have a hardware problem.  After you get a case opened, you can try one thing to see if you can get HA active again without having to re-image the current Slave:

    1. In the Master, set HA to "Off" - the Slave will do a factory reset and shut down.
    2. Set HA back to "Hot-Standby."
    3. Power up the Slave.  It should rejoin, sync and return to status READY.

    Cheers - Bob

     
    Sophos UTM Community Moderator
    Sophos Certified Architect - UTM
    Sophos Certified Engineer - XG
    Gold Solution Partner since 2005
    MediaSoft, Inc. USA
  • Bob,

    thank you for your reply. Since I am running some UTMs for several years I do know what a normal HA up2date process looks like and I know that this HA cluster has no hardware problems at all, too. :)

    But thank you anyway for your input.

     

    I resolved it some minutes ago (at the weekend I can risk downtimes) by myself:

    - deleted up2date packages on node1 & 2 (node2 is slave and the only one at 9.506)

    - fired up update process on node1 at command line

    - rebooted node1

    - slave (still in reserved mode) did not take over

    - node 1 came online again

    - HA came up by itself

     

    So if anyone gets stuck in the same thing - give it a try but you get a downtime during that procedure.

     

    Cheers, Chris

  • Yes - if the Master was not yet at 9.506, your solution is exactly correct, Chris.

    For others, if the Slave is not in Reserved mode, you will want to power it down and then back up again after the Master has finished its Up2Dating and is back up.

    I appreciate that Sophos' Support policies are different for Europe and North America.

    Cheers - Bob

     
    Sophos UTM Community Moderator
    Sophos Certified Architect - UTM
    Sophos Certified Engineer - XG
    Gold Solution Partner since 2005
    MediaSoft, Inc. USA
  • BAlfson said:

    I appreciate that Sophos' Support policies are different for Europe and North America.

     

     

    I am just curios: Where is the difference? The last time a opened up a support call is 2 years ago so I don´t know...

     

    Cheers - Chris

Reply
  • BAlfson said:

    I appreciate that Sophos' Support policies are different for Europe and North America.

     

     

    I am just curios: Where is the difference? The last time a opened up a support call is 2 years ago so I don´t know...

     

    Cheers - Chris

Children
  • In general, in NA, customers with Standard Support contact their reseller to get a case open directly with Sophos and we don't charge extra to do that.

    Cheers - Bob

     
    Sophos UTM Community Moderator
    Sophos Certified Architect - UTM
    Sophos Certified Engineer - XG
    Gold Solution Partner since 2005
    MediaSoft, Inc. USA
  • Same here in Europe.

    Cheers, Chris

  • There must have been a change because I have worked with people in Germany over the years.  If that's the case, I would want to have Sophos look at the logs just to confirm that there wasn't a hardware glitch at the bottom of this.

    Cheers - Bob

     
    Sophos UTM Community Moderator
    Sophos Certified Architect - UTM
    Sophos Certified Engineer - XG
    Gold Solution Partner since 2005
    MediaSoft, Inc. USA