Skip to main content
pj255
New Member
November 25, 2014
Solved

FortiGate Active-Standby Cluster - Seperate Management IP addess

  • November 25, 2014
  • 8 replies
  • 61707 views

Hi,

 

We're running a pair of 1000C's in A-P (v5.0,build3608 (GA Patch 7)).

 

We currently manage both FW's using MGMT1 with one dedicated IP.

 

Does anyone how to give the primary and secondary separate dedicated MGMT IP's ?

 

I'd like to use MGMT 1 on Primary and MGMT 2 on secondary - each with a different IP address.

 

Thanks,

PJ

    Best answer by kritt

    Hello mark9885,

     

    I 've had the same error.

    In my case, this error has been resolved by deleting :

    - static route associated to mgmt

    - source-ip setting in the syslog server config

     

    Maybe that the error could appear if the mgmt interface is part of  firewall policies, this error could appear to.

     

    I guess that the interface has not to be part of a specific configuration before to be used as reserved management interface.

     

     

    8 replies

    Jeff_FTNT
    Staff
    Staff
    November 25, 2014

    You may try enable "ha-magmt-status " on both Master and Slave, set up different IP, you can manage Master and Slave with different IP.

    Config sys ha

    set ha-mgmt-status     enable set ha-mgmt-interface   xx set ha-mgmt-interface-gateway x.x.x.x

    end

    pj255
    pj255Author
    New Member
    November 27, 2014

    Thanks Jeff!!!

     

    il get a change scheduled and test this out.

     

     

     

    Jeff_FTNT wrote:

    You may try enable "ha-magmt-status " on both Master and Slave, set up different IP, you can manage Master and Slave with different IP.

    Config sys ha

    set ha-mgmt-status     enable set ha-mgmt-interface   xx set ha-mgmt-interface-gateway x.x.x.x

    end

    Jeff_FTNT
    Staff
    Staff
    December 1, 2014

    Hi Fullmoon,

    HA have a option:set uninterruptable-upgrade {disable | enable}

    You may try it in LAB with "set uninterruptable-upgrade enable". Before upgrade, it is better to back up setting.

    For FSSO, no need any change, the user information will get from Windows AD server after upgrade. Hope it have some help, thanks.

    Fullmoon
    New Member
    December 4, 2014

    Jeff_FTNT wrote:

    Hi Fullmoon,

    HA have a option:set uninterruptable-upgrade {disable | enable}

    You may try it in LAB with "set uninterruptable-upgrade enable". Before upgrade, it is better to back up setting.

    For FSSO, no need any change, the user information will get from Windows AD server after upgrade. Hope it have some help, thanks.

    I Jeff thank you for your reply. I guess by default "set uninterruptable-upgrade" was set to Enable. In my lab tried to upgrade my HA A-A couple of times with same effect.

    Pls correct me if Im wrong if these procedure are correct, In my lab, updating my HA A-A thru GUI and to get a better picture whats going on behind my slave unit connect my console cable and seems updating works fine,I had a computer continuesly pinging to fortigate local ip and www.yahoo.com to check if theres any lose or rto's. Wait for a couple of minutes, then upgrade the firmware of the Master unit thru GUI, heres what I found out, seems Slave unit takes time to kick-in while the Master unit is the process of firmware upgrade. And I got 5-10 rto's before everything backs to normal.

     

    Heres my HA settings for better picture

    system hagroup-id : 0 group-name : FGT-HA mode : a-a password : * hbdev : "internal1" 50 session-sync-dev : route-ttl : 10 route-wait : 0 route-hold : 10 sync-config : enable encryption : disable authentication : disable hb-interval : 2 hb-lost-threshold : 6 helo-holddown : 20 arps : 5 arps-interval : 8 session-pickup : enable session-pickup-connectionless: disable session-pickup-delay: disable update-all-session-timer: disable session-sync-daemon-number: 1 link-failed-signal : disable uninterruptable-upgrade: enable ha-mgmt-status : enable ha-mgmt-interface : internal5 ha-mgmt-interface-gateway: 0.0.0.0 ha-eth-type : 8890 hc-eth-type : 8891 l2ep-eth-type : 8893 ha-uptime-diff-margin: 300 vcluster2 : disable vcluster-id : 1 override : disable priority : 128 schedule : round-robin monitor : "internal4" "wan1" pingserver-monitor-interface: pingserver-failover-threshold: 0 pingserver-flip-timeout: 60 vdom : "root" load-balance-all : disable

     

     

     

     

     

     

    Jeff_FTNT
    Staff
    Staff
    December 4, 2014

    "then upgrade the firmware of the Master unit thru GUI, heres what I found out, seems Slave unit takes time to kick-in while the Master unit is the process of firmware upgrade."

     

    Did you upgrade from Slave GUI firstly ? 

    For upgrade, we just login from Master GUI and do upgrade. Master will send image to Slave and upgrade, Master wait for Slave finish upgrade, then upgrade itself. Thanks.

    Fullmoon
    New Member
    December 4, 2014

    Hi Jeff

     

    The working mechanism of 'uninterruptable-upgrade' if ENABLED is as follows - I log into the webgui  - the webgui will show me the master  - I uploaded the firmware to the master  - the master will transfer the firmware to the slave using the heartbeat cable  - the slave will perform the firmware upgrade first  - during the slave upgrade, there will be no downtime because the master is still up  - when the slave is up,---------- The Master Unit didnt perform firmware upgrade this is the portion where I am lost. I can't comprehend why my Master unit didn't performing firmware upgrade. After the Slave successfully upgraded and totally UP, I waited for almost 10 mins or more to check what would happen next but it seems in GUI Master didn't initiate firmware upgrade until i forced to upload manually the firmware to Master unit. Can you spot my error why my Master didn't update its firmware after Slave successfully updated?

     

    thanks

    Jeff_FTNT
    Staff
    Staff
    December 5, 2014

    It look some thing is wrong.

    After Slave upgrade, it will new master, old Master will upgrade image.

     

    Before you do upgrade, make sure HA is synchronize well, check it with CLI:dia sys ha showcsum , both master and slave should have same checksum.Thanks.

    Fullmoon
    New Member
    December 6, 2014

     

    Hi Jeff,

    heres the outcome. thank you

    diagnose sys ha showcsum 
    is_manage_master()=1, is_root_master()=1
    debugzone
    global: 1b 2f 63 20 de 0b 2f 53 5c 73 b6 05 7a 52 ef d9 
    root: 78 02 f6 24 d1 d0 df 50 10 4c a2 84 d3 0f d2 a6 
    all: d9 b0 e0 10 5b 63 f5 e8 11 a4 9f b3 86 26 02 ba 
     
    checksum
    global: 1b 2f 63 20 de 0b 2f 53 5c 73 b6 05 7a 52 ef d9 
    root: 78 02 f6 24 d1 d0 df 50 10 4c a2 84 d3 0f d2 a6 
    all: d9 b0 e0 10 5b 63 f5 e8 11 a4 9f b3 86 26 02 ba 
    Jeff_FTNT
    Staff
    Staff
    December 8, 2014

    Hi Fullmoon,

    From output, your HA is synchronized. Thanks.

    Fullmoon
    New Member
    December 9, 2014

    hi Jeff, my HA is now working. From ver 4.0 MR3 P18 I jumped in to Ver 5.0.4 not 5.0.0. Im not quite sure why 5.0.0 is not working well in HA firmware upgarde.

    Additional verification do I need to perform the same action if Im going to upgrade HA in A-P mode? Thanks

    Jeff_FTNT
    Staff
    Staff
    December 9, 2014

    Upgrade from v4.3  to v5.0 need special step, can not jump directly.

    I remember it need upgrade to the newest v4.3 , then upgrade to v5.0. You may find some DOC on support site.

     

    Like:http://www.google.ca/url?sa=t&rct=j&q=&esrc=s&source=web&cd=2&cad=rja&uact=8&ved=0CCUQFjAB&url=http%3A%2F%2Fdocs.fortinet.com%2Fd%2Fsupported-upgrade-paths-1%2Fdownload&ei=4iqHVLHPMZPnoAT27oHoAQ&usg=AFQjCNHYbWKPzWFRxK9bssPBOBow-uczzw&sig2=B-JtzV41h_sjJW_ipbGcWg&bvm=bv.81449611,d.cGU

     

    Thanks.

    Toshi_Esumi
    SuperUser
    SuperUser
    December 9, 2014

    At least we didn't have any problem upgrading from v4.3.14 directly to v5.0.7. From there we could go to v5.0.9.

    Since the last SSL bug, I recommend v5.0.7 or above. Also upgrading to v5.0 will throw some overlapped subnets/duplicated entries, and so on. All should be in config-error-log.

    ede_pfau
    SuperUser
    SuperUser
    September 18, 2015

    In FortiOS v5.2.3 (at least), these settings can be made directly in the GUI, Config > HA. Really straightforward.

    Petzocles
    New Member
    March 24, 2016

    I have done:

     

    set ha-mgmt-status     enable  set ha-mgmt-interface   xx  set ha-mgmt-interface-gateway x.x.x.x

     

    but now I see that HA synchronises mgmt IP's 

    so now both nodes have the same mgmt IP and now I even can manage none of the nodes via network. not even in the segment itself! 

     

    how can I make sure he does not sync the IP address!

     

     

    rough
    New Member
    March 28, 2016

    You should config reserved management interface first.

    To configure the reserved management interface - web-based manager 1. Go to System > Config > HA. 2. Edit the primary unit. 3. Select Reserve Management Port for Cluster Member and select portxxx. 4. Select OK.

    You can also get more information here: http://docs.fortinet.com/...88/fortigate-ha-50.pdf