Rocket iCluster

Hello

I'm running iCluster 9.1.1 in IBM i 7.4 and have a need to change IP address used by iCluster in primary and backup nodes from a.b.c.9 to a.b.c.10 at primary and x.y.z.9 to x.y.z.10 at backup node.   I find that the instruction to change IP addresses in manual of iCluster 9.1.0 does not work but my Google search leads me to iCluster 9.2.0 instruction on how to do this change at  https://docs.rocketsoftware.com/bundle/icluster_ug_920/page/hft1701835840938.html  which provides new action steps and a new command DMCHGNODIP for use.  I follow the instruction and found it does not work.   I suspect that action steps for 9.1.1 may be different from 9.2.

What I do is

1. Log on to primary node and end all replication groups and  ENDSBS XDMCLUSTER in both nodes as instructed in the URL above
2. Run DMCHGNODIP for Backup node first and there is a message that the Backup node must be active
3. So, I restart XDMCLUSTER subsystem for both nodes back
4. I then run  DMCHGNODIP for Backup node definition and see the message that the node must not be active which is strange
5. I end Backup node leaving Primary node active and I can successfully run DMCHGNODIP for Backup node definition and I display it to confirm the change but I notice this is opposite to the instruction for 9.2 and make me feel errie 
6. I run DMCHGNODIP for Primary node definition and a message appears that the node must be inactive so I end primary node and run DMCHGNODIP again and a new message appears saying that "The current node is not active in the cluster."  The message asks thenode to be active which blatantly contradicts to its previous message.    At this point I think the instruction for 9.1.0 and 9.2.0 do not work for 9.1.1 but I cannot find instruction for 9.1.1 and am now lost.   Or is this a bug for DMCHGNODIP in 9.1.1? 
   
   

Now I'm in a state that I cannot change the new IP address for primary node definition.    I call my iCluster BP and am suggested that he heard there is an alternative way when I end XDMCLUSTER subsystem and use SQL to change new IP address for both primary and backup nodes in a table used by iCluster manually in both systems.  So, I post this question to ask if any one of you know of this alternative way and provide instruction for me to follow?   (I know enough how to use SQL in DB2 for i.  I normally use Run SQL Scripts tool from IBM i ACS).

I thank you in advance for your kind help on this matter.  

------------------------------
Satid Singkorapoom
IBM i SME
Rocket Forum Shared Account
------------------------------

Now I know the issue.  When I first successfully changed ip address for backup node definition, I restarted XDMCLUSTER subsystem which means backup node then ran on the new ip address.  Thereafter, the reason DMCHGNODIP failed on primary node definition was that there is a firewall blocking any communication between backup node's new ip and primary node's old ip.  (The firewall already allows in advance comm between the new ip pair.)   I also cannot start the backup node by this same firewall blockage.

Asking for firewall policy to allow this temp operation must pass a red tape process that can take long.  So, I see that manual update of the new ip addresses with SQL to iCluster's table in both IBM i PRD and DR can be a better solution to this issue and I hope a kind soul would understand and provide me with an instruction on how to accomplish this.

Thanks. 

------------------------------
Satid Singkorapoom
IBM i SME
Rocket Forum Shared Account

Original Message:
Sent: 01-26-2024 07:50
From: Satid Singkorapoom
Subject: Help need - Changing new IP addresses for both node definitions is only half successful

Hello

I'm running iCluster 9.1.1 in IBM i 7.4 and have a need to change IP address used by iCluster in primary and backup nodes from a.b.c.9 to a.b.c.10 at primary and x.y.z.9 to x.y.z.10 at backup node.   I find that the instruction to change IP addresses in manual of iCluster 9.1.0 does not work but my Google search leads me to iCluster 9.2.0 instruction on how to do this change at  https://docs.rocketsoftware.com/bundle/icluster_ug_920/page/hft1701835840938.html  which provides new action steps and a new command DMCHGNODIP for use.  I follow the instruction and found it does not work.   I suspect that action steps for 9.1.1 may be different from 9.2.

What I do is

1. Log on to primary node and end all replication groups and  ENDSBS XDMCLUSTER in both nodes as instructed in the URL above
2. Run DMCHGNODIP for Backup node first and there is a message that the Backup node must be active
3. So, I restart XDMCLUSTER subsystem for both nodes back
4. I then run  DMCHGNODIP for Backup node definition and see the message that the node must not be active which is strange
5. I end Backup node leaving Primary node active and I can successfully run DMCHGNODIP for Backup node definition and I display it to confirm the change but I notice this is opposite to the instruction for 9.2 and make me feel errie 
6. I run DMCHGNODIP for Primary node definition and a message appears that the node must be inactive so I end primary node and run DMCHGNODIP again and a new message appears saying that "The current node is not active in the cluster."  The message asks thenode to be active which blatantly contradicts to its previous message.    At this point I think the instruction for 9.1.0 and 9.2.0 do not work for 9.1.1 but I cannot find instruction for 9.1.1 and am now lost.   Or is this a bug for DMCHGNODIP in 9.1.1? 
   
   

Now I'm in a state that I cannot change the new IP address for primary node definition.    I call my iCluster BP and am suggested that he heard there is an alternative way when I end XDMCLUSTER subsystem and use SQL to change new IP address for both primary and backup nodes in a table used by iCluster manually in both systems.  So, I post this question to ask if any one of you know of this alternative way and provide instruction for me to follow?   (I know enough how to use SQL in DB2 for i.  I normally use Run SQL Scripts tool from IBM i ACS).

I thank you in advance for your kind help on this matter.  

------------------------------
Satid Singkorapoom
IBM i SME
Rocket Forum Shared Account
------------------------------

I realize I had iCluster support account and used it to open a case and got a response very quickly on using SQL to manually update iCluster table for the new ip.  All works now. 

Thanks. 

Now I know the issue.  When I first successfully changed ip address for backup node definition, I restarted XDMCLUSTER subsystem which means backup node then ran on the new ip address.  Thereafter, the reason DMCHGNODIP failed on primary node definition was that there is a firewall blocking any communication between backup node's new ip and primary node's old ip.  (The firewall already allows in advance comm between the new ip pair.)   I also cannot start the backup node by this same firewall blockage.

Asking for firewall policy to allow this temp operation must pass a red tape process that can take long.  So, I see that manual update of the new ip addresses with SQL to iCluster's table in both IBM i PRD and DR can be a better solution to this issue and I hope a kind soul would understand and provide me with an instruction on how to accomplish this.

Thanks. 

------------------------------
Satid Singkorapoom
IBM i SME
Rocket Forum Shared Account

Original Message:
Sent: 01-26-2024 07:50
From: Satid Singkorapoom
Subject: Help need - Changing new IP addresses for both node definitions is only half successful

Hello

I'm running iCluster 9.1.1 in IBM i 7.4 and have a need to change IP address used by iCluster in primary and backup nodes from a.b.c.9 to a.b.c.10 at primary and x.y.z.9 to x.y.z.10 at backup node.   I find that the instruction to change IP addresses in manual of iCluster 9.1.0 does not work but my Google search leads me to iCluster 9.2.0 instruction on how to do this change at  https://docs.rocketsoftware.com/bundle/icluster_ug_920/page/hft1701835840938.html  which provides new action steps and a new command DMCHGNODIP for use.  I follow the instruction and found it does not work.   I suspect that action steps for 9.1.1 may be different from 9.2.

What I do is

1. Log on to primary node and end all replication groups and  ENDSBS XDMCLUSTER in both nodes as instructed in the URL above
2. Run DMCHGNODIP for Backup node first and there is a message that the Backup node must be active
3. So, I restart XDMCLUSTER subsystem for both nodes back
4. I then run  DMCHGNODIP for Backup node definition and see the message that the node must not be active which is strange
5. I end Backup node leaving Primary node active and I can successfully run DMCHGNODIP for Backup node definition and I display it to confirm the change but I notice this is opposite to the instruction for 9.2 and make me feel errie 
6. I run DMCHGNODIP for Primary node definition and a message appears that the node must be inactive so I end primary node and run DMCHGNODIP again and a new message appears saying that "The current node is not active in the cluster."  The message asks thenode to be active which blatantly contradicts to its previous message.    At this point I think the instruction for 9.1.0 and 9.2.0 do not work for 9.1.1 but I cannot find instruction for 9.1.1 and am now lost.   Or is this a bug for DMCHGNODIP in 9.1.1? 
   
   

Now I'm in a state that I cannot change the new IP address for primary node definition.    I call my iCluster BP and am suggested that he heard there is an alternative way when I end XDMCLUSTER subsystem and use SQL to change new IP address for both primary and backup nodes in a table used by iCluster manually in both systems.  So, I post this question to ask if any one of you know of this alternative way and provide instruction for me to follow?   (I know enough how to use SQL in DB2 for i.  I normally use Run SQL Scripts tool from IBM i ACS).

I thank you in advance for your kind help on this matter.  

------------------------------
Satid Singkorapoom
IBM i SME
Rocket Forum Shared Account
------------------------------