RHEL Cluster -- MQ as HA cluster resource

5. RE: RHEL Cluster -- MQ as HA cluster resource

Like

RAJESH VERMA

Posted Tue July 05, 2022 04:40 PM

Hi Morag,
Thank you very much sharing the information.

Thank you,
With regards.
Rajesh Verma

------------------------------
RAJESH VERMA
------------------------------

Original Message

6. RE: RHEL Cluster -- MQ as HA cluster resource

Like

Jamie Squibb

Posted Wed July 06, 2022 03:57 AM

Thank you for replying Morag.

If you require any further information Rajesh, then please let us know.

Kind regards,
Jamie

------------------------------
Jamie Squibb
IBM MQ Development
------------------------------

Original Message

7. RE: RHEL Cluster -- MQ as HA cluster resource

Like

RAJESH VERMA

Posted Wed July 06, 2022 01:36 PM

Hi Morag and Jamie,
Thank you both for being so kind. I have few more questions;
1) I have two MQ queuemanager, Dev and Prod. Do I really need to use three NICs for MQ -DEV -RDQM setup. I had setup this with one NIC on each of three Linux VM, do you think I needed three NICs for DEV MQ RDQM as well?
2) Do I need to put additional configuration into this newly setup DEV RDQM after importing the configuration data from AIX server using dmpmqcfg to be able to connect using MQ Explorer to queue manager in RDQM env.

I would definitely go with three VM - each of these with three NICs for the Production MQ RDQM environment.

I think for now I have only these questions, I will ask for more help as I continue to make progress to migrate both MQ environments to RDQM.

Thank you,
Regards,
Rajesh

------------------------------
RAJESH VERMA
------------------------------

Original Message

8. RE: RHEL Cluster -- MQ as HA cluster resource

Like

Jamie Squibb

Posted Wed July 06, 2022 04:53 PM

Hi Rajesh,

> Do I really need to use three NICs for MQ -DEV -RDQM setup?

You can configure an RDQM HA group to use either one, two, or three IP addresses. An RDQM HA group has three logical interfaces (network connections) between the RDQM nodes. Two of the interfaces (primary and alternate) are used by Pacemaker to exchange heartbeats and coordinate configuration changes. The third interface is used by DRBD to replicate data between the nodes.

- If you configure a single IP address, then all three interfaces share it, which simplifies the RDQM configuration.
- If you configure two IP addresses, then one IP address is used by Pacemaker (primary and alternate) and the other IP address is used by DRBD.
- If you configure three IP addresses, then each interface has a dedicated IP address.

The number of IP addresses you choose to use depends on the environment you are using and the resilience that you require. For maximum resiliency you might choose to use three IP addresses, ensure that each IP address maps to a dedicated network card (NIC), and ensure that the network infrastructure used by the connections (switches, routers, etc.) is not shared. This configuration helps to remove single points of failure and minimise the risk of the nodes losing connectivity. However, there are other ways to achieve resiliency. For example, you might use techniques such as link aggregation to provide a single resilient network connection instead that is based on multiple physical links. This alternative approach achieves resiliency at the system/network layer, which can then be transparent to RDQM. In virtualised environments it might also be difficult to map virtual network interfaces to physical network ports, and in virtualised development environments all nodes might be hosted on the same physical server anyway.

RDQM is functionally the same for each option, so you do not need to use the same number of IP addresses in each staging environment (dev, QA, prod, etc.) if they do not require the same level of resiliency. For example, it would be common to have a greater tolerance for outages in a development environment than in production. You just need to decide whether any difference between environments is acceptable to your confidence in promoting from one environment to another.

In summary, the short answer to your question is no, you don't need to use three NICs in your development environment if you don't want to. The above longer answer provides some background for what you might want to consider when deciding if this is the correct choice for you or not.

> Do I need to put additional configuration into this newly setup DEV RDQM after importing the configuration data from AIX server using dmpmqcfg to be able to connect using MQ Explorer to queue manager in RDQM env?

I would recommend that you review MQ object definitions that refer to network connections, such as listeners and channels, to ensure they correctly reflect your RDQM setup. For example, if you are using a floating IP address you probably want to ensure that your listener(s) bind to the floating IP. This helps to ensure that applications are connecting to the floating IP instead of to a node's static IP address by mistake, which would not work after a HA failover to different node. Some specific items to check are:

- Local IP addresses used by listeners (IPADDR attribute).
- Local IP addresses used by channels (LOCLADDR attribute).
- Local IP addresses defined for cluster-receiver channels (CONNAME attribute).
- Local IP addresses defined for communication information objects (GRPADDR attribute).
- Channel authentication rules that might reference specific hostnames, IP addresses, or subnets.

You might also want to review your security configuration to determine if you need to adjust anything for user authentication and authorisation, especially if you are using local operating system userids that might need to be defined on each RDQM node.

You'll also need to review the configuration of applications and any remote queue managers that connect in to your new RDQM queue manager to ensure they no longer attempt to connect to your old AIX server.

I hope this helps.

Kind regards,
Jamie

------------------------------
Jamie Squibb
IBM MQ Development
------------------------------

Original Message

9. RE: RHEL Cluster -- MQ as HA cluster resource

Like

RAJESH VERMA

Posted Wed July 06, 2022 09:49 PM

Hi Jamie,

I really appreciate that you took some time to answer and explain with lot of details.

Thank you very much!

With regards,

Rajesh

Original Message

10. RE: RHEL Cluster -- MQ as HA cluster resource

Like

RAJESH VERMA

Posted Fri July 08, 2022 12:23 PM

Hi Jamie and Morag,
I am not an MQ administator, would you please guide me how can I now connect to the defined RDQM-Queue Manager using MQ Explorer or any other tool.

Thank you,
Regards,
Rajesh

------------------------------
RAJESH VERMA
------------------------------

Original Message

11. RE: RHEL Cluster -- MQ as HA cluster resource

Like

RAJESH VERMA

Posted Mon July 11, 2022 11:43 AM

Hi Jamie and Morag,
As I am just system administrator with very limited knowledge about MQ, would you please guide me how can I now connect to this newly defined RDQM-Queue Manager using MQ Explorer or any other tool.
As I compared with my previous dev environment, it seems like RDQM does not run the listener program and due to I was not able to connect using MQ Explorer, the moment I run that I was able to connect to it.

But I believe that should run it alongwith the MQ when RDQM starts. Anyway I am not going to go this route to migrate any of the MQ instances (dev and prod). Now I will try to attempt migrate MQ on openshift and please be kind to let me know for any kind of procedure/steps.

Thank you very much.

Regards,
Rajesh

Original Message

12. RE: RHEL Cluster -- MQ as HA cluster resource

Like

Luc-Michel Demey

IBM Champion

Posted Mon July 11, 2022 11:54 AM

Hello Rajesh,

A RDQM Queue Manager is just a regular Queue Manager, there is nothing special about the listener.
You need to create a listener via an MQSC command or something, and make sure it is definied with CONTROL(QMGR) so that it starts with the Queue Manager.
If the Queue Manager switches to another node, it is restarted on that node and the listener starts as usual.
Furthermore, as the IP address of the Queue Manager is a VIP moved by the RDQM infrastructure, there are no additional parameters to set in MQ Explorer. The configuration is the same as for a standalone Queue Manager.

HTH, LMD.

------------------------------
Luc-Michel Demey
DEMEY CONSULTING
lmd@demey-consulting.fr
#IBMChampion
------------------------------

Original Message

13. RE: RHEL Cluster -- MQ as HA cluster resource

Like

RAJESH VERMA

Posted Wed July 13, 2022 04:18 PM

Hi Luc,
Would you please give some details as how do I define listener which automatically starts or move when RDQM starts the queue manager or move to another node.

Thank you,

Regards,
Rajesh

------------------------------
RAJESH VERMA
------------------------------

Original Message

14. RE: RHEL Cluster -- MQ as HA cluster resource

Like

Morag Hughson

IBM Champion

Posted Wed July 13, 2022 04:23 PM

Hi Rajesh,

As @Luc-Michel Demey said in his response before, you need to define the listener object with CONTROL(QMGR) in order for it to start automatically with the queue manager. This means the full command is as follows:-

DEFINE LISTENER(MY.FLOATING.LSTR) TRPTYPE(TCP) IPADDR('floating-ip-addr') PORT(port-number) CONTROL(QMGR)

Then the first time only, you also need to start this listener. Subsequently it will be started whenever the queue manager starts.

START LISTENER(MY.FLOATING.LSTR)

Cheers,
Morag

------------------------------
Morag Hughson
MQ Technical Education Specialist
MQGem Software Limited
Website: https://www.mqgem.com
------------------------------

Original Message

15. RE: RHEL Cluster -- MQ as HA cluster resource

Like

RAJESH VERMA

Posted Wed July 13, 2022 04:52 PM

Hi Morag,
The first command is giving me an error;

mqm@alphamqdev1:~ $ DEFINE LISTENER(MY.FLOATING.LSTR) TRPTYPE(TCP) IPADDR('192.0.1.58') PORT(1414) CONTROL(QM.MQT)
ksh: syntax error: unexpected '('
1|mqm@alphamqdev1:~ $

Thanks.
Regards,
Rajesh

------------------------------
RAJESH VERMA
------------------------------

Original Message

Original Message:
Sent: Wed July 13, 2022 04:23 PM
From: Morag Hughson
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Rajesh,

As @Luc-Michel Demey said in his response before, you need to define the listener object with CONTROL(QMGR) in order for it to start automatically with the queue manager. This means the full command is as follows:-

DEFINE LISTENER(MY.FLOATING.LSTR) TRPTYPE(TCP) IPADDR('floating-ip-addr') PORT(port-number) CONTROL(QMGR)

Then the first time only, you also need to start this listener. Subsequently it will be started whenever the queue manager starts.

START LISTENER(MY.FLOATING.LSTR)

Cheers,
Morag

------------------------------
Morag Hughson
MQ Technical Education Specialist
MQGem Software Limited
Website: https://www.mqgem.com

Original Message:
Sent: Wed July 13, 2022 04:17 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Luc,
Would you please give some details as how do I define listener which automatically starts or move when RDQM starts the queue manager or move to another node.

Thank you,

Regards,
Rajesh

------------------------------
RAJESH VERMA

Original Message:
Sent: Mon July 11, 2022 11:53 AM
From: Luc-Michel Demey
Subject: RHEL Cluster -- MQ as HA cluster resource

Hello Rajesh,

A RDQM Queue Manager is just a regular Queue Manager, there is nothing special about the listener.
You need to create a listener via an MQSC command or something, and make sure it is definied with CONTROL(QMGR) so that it starts with the Queue Manager.
If the Queue Manager switches to another node, it is restarted on that node and the listener starts as usual.
Furthermore, as the IP address of the Queue Manager is a VIP moved by the RDQM infrastructure, there are no additional parameters to set in MQ Explorer. The configuration is the same as for a standalone Queue Manager.

HTH, LMD.

------------------------------
Luc-Michel Demey
DEMEY CONSULTING
lmd@demey-consulting.fr
#IBMChampion

Original Message:
Sent: Mon July 11, 2022 11:42 AM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Jamie and Morag,
As I am just system administrator with very limited knowledge about MQ, would you please guide me how can I now connect to this newly defined RDQM-Queue Manager using MQ Explorer or any other tool.
As I compared with my previous dev environment, it seems like RDQM does not run the listener program and due to I was not able to connect using MQ Explorer, the moment I run that I was able to connect to it.

But I believe that should run it alongwith the MQ when RDQM starts. Anyway I am not going to go this route to migrate any of the MQ instances (dev and prod). Now I will try to attempt migrate MQ on openshift and please be kind to let me know for any kind of procedure/steps.

Thank you very much.

Regards,
Rajesh

Original Message:
Sent: 7/6/2022 9:49:00 PM
From: RAJESH VERMA
Subject: RE: RHEL Cluster -- MQ as HA cluster resource

Hi Jamie,

I really appreciate that you took some time to answer and explain with lot of details.

Thank you very much!

With regards,

Rajesh

Original Message:
Sent: 7/6/2022 4:53:00 PM
From: Jamie Squibb
Subject: RE: RHEL Cluster -- MQ as HA cluster resource

Hi Rajesh,

> Do I really need to use three NICs for MQ -DEV -RDQM setup?

You can configure an RDQM HA group to use either one, two, or three IP addresses. An RDQM HA group has three logical interfaces (network connections) between the RDQM nodes. Two of the interfaces (primary and alternate) are used by Pacemaker to exchange heartbeats and coordinate configuration changes. The third interface is used by DRBD to replicate data between the nodes.

- If you configure a single IP address, then all three interfaces share it, which simplifies the RDQM configuration.
- If you configure two IP addresses, then one IP address is used by Pacemaker (primary and alternate) and the other IP address is used by DRBD.
- If you configure three IP addresses, then each interface has a dedicated IP address.

The number of IP addresses you choose to use depends on the environment you are using and the resilience that you require. For maximum resiliency you might choose to use three IP addresses, ensure that each IP address maps to a dedicated network card (NIC), and ensure that the network infrastructure used by the connections (switches, routers, etc.) is not shared. This configuration helps to remove single points of failure and minimise the risk of the nodes losing connectivity. However, there are other ways to achieve resiliency. For example, you might use techniques such as link aggregation to provide a single resilient network connection instead that is based on multiple physical links. This alternative approach achieves resiliency at the system/network layer, which can then be transparent to RDQM. In virtualised environments it might also be difficult to map virtual network interfaces to physical network ports, and in virtualised development environments all nodes might be hosted on the same physical server anyway.

RDQM is functionally the same for each option, so you do not need to use the same number of IP addresses in each staging environment (dev, QA, prod, etc.) if they do not require the same level of resiliency. For example, it would be common to have a greater tolerance for outages in a development environment than in production. You just need to decide whether any difference between environments is acceptable to your confidence in promoting from one environment to another.

In summary, the short answer to your question is no, you don't need to use three NICs in your development environment if you don't want to. The above longer answer provides some background for what you might want to consider when deciding if this is the correct choice for you or not.

> Do I need to put additional configuration into this newly setup DEV RDQM after importing the configuration data from AIX server using dmpmqcfg to be able to connect using MQ Explorer to queue manager in RDQM env?

I would recommend that you review MQ object definitions that refer to network connections, such as listeners and channels, to ensure they correctly reflect your RDQM setup. For example, if you are using a floating IP address you probably want to ensure that your listener(s) bind to the floating IP. This helps to ensure that applications are connecting to the floating IP instead of to a node's static IP address by mistake, which would not work after a HA failover to different node. Some specific items to check are:

- Local IP addresses used by listeners (IPADDR attribute).
- Local IP addresses used by channels (LOCLADDR attribute).
- Local IP addresses defined for cluster-receiver channels (CONNAME attribute).
- Local IP addresses defined for communication information objects (GRPADDR attribute).
- Channel authentication rules that might reference specific hostnames, IP addresses, or subnets.

You might also want to review your security configuration to determine if you need to adjust anything for user authentication and authorisation, especially if you are using local operating system userids that might need to be defined on each RDQM node.

You'll also need to review the configuration of applications and any remote queue managers that connect in to your new RDQM queue manager to ensure they no longer attempt to connect to your old AIX server.

I hope this helps.

Kind regards,
Jamie

------------------------------
Jamie Squibb
IBM MQ Development

Original Message:
Sent: Wed July 06, 2022 01:36 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Morag and Jamie,
Thank you both for being so kind. I have few more questions;
1) I have two MQ queuemanager, Dev and Prod. Do I really need to use three NICs for MQ -DEV -RDQM setup. I had setup this with one NIC on each of three Linux VM, do you think I needed three NICs for DEV MQ RDQM as well?
2) Do I need to put additional configuration into this newly setup DEV RDQM after importing the configuration data from AIX server using dmpmqcfg to be able to connect using MQ Explorer to queue manager in RDQM env.

I would definitely go with three VM - each of these with three NICs for the Production MQ RDQM environment.

I think for now I have only these questions, I will ask for more help as I continue to make progress to migrate both MQ environments to RDQM.

Thank you,
Regards,
Rajesh

------------------------------
RAJESH VERMA

Original Message:
Sent: Wed July 06, 2022 03:57 AM
From: Jamie Squibb
Subject: RHEL Cluster -- MQ as HA cluster resource

Thank you for replying Morag.

If you require any further information Rajesh, then please let us know.

Kind regards,
Jamie

------------------------------
Jamie Squibb
IBM MQ Development

Original Message:
Sent: Tue July 05, 2022 04:40 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Morag,
Thank you very much sharing the information.

Thank you,
With regards.
Rajesh Verma

------------------------------
RAJESH VERMA

Original Message:
Sent: Tue July 05, 2022 04:16 PM
From: Morag Hughson
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Rajesh,

You can use dmpmqcfg to save your MQ Configuration on the original AIX Queue Manager, and then to restore the configuration on the new RHEL Queue Manager.

If you want to copy message data as well, something like QLOAD generic unload/reload might be of interest.

Cheers,
Morag

------------------------------
Morag Hughson
MQ Technical Education Specialist
MQGem Software Limited
Website: https://www.mqgem.com

Original Message:
Sent: Tue July 05, 2022 02:07 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Jamie,
I am trying to get some help for past couple of weeks, then I see your reply to Ishara. I am working to migrate my current estate of MQ from AIX to RDQM. I have almost setup everything and looking good. Now I want to know how can I bring my current MQ configuration along with data (if possible) from AIX to this new RDQM cluster. I request you to please share some insight on this.

Thank you very much.
Regards,
Rajesh

------------------------------
RAJESH VERMA

Original Message:
Sent: Wed October 07, 2020 04:31 AM
From: Jamie Squibb
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Ishara,

I would suggest you review the Replicated Data Queue Manager (RDQM) feature that is available for RHEL 7 and RHEL 8. RDQM is a high availability and disaster recovery capability of MQ Advanced that is provided by IBM. It is built using DRBD for disk replication and Pacemaker for HA cluster management. RDQM manages the Pacemaker configuration for you.

See https://www.ibm.com/support/knowledgecenter/en/SSFKSJ_9.2.0/com.ibm.mq.con.doc/q130280_.htm

Kind regards,
Jamie

------------------------------
Jamie Squibb

Original Message:
Sent: Mon October 05, 2020 07:02 AM
From: Ishara Pathirana
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi all.

I want to get some info about how to add MQ as a cluster resource. In a Redhat Pacemaker cluster

thanks.

------------------------------
Ishara Pathirana
------------------------------

16. RE: RHEL Cluster -- MQ as HA cluster resource

Like

Morag Hughson

IBM Champion

Posted Wed July 13, 2022 04:53 PM

It's an MQSC command. You must issue it using the runmqsc tool (or an equivalent).

Cheers,
Morag

------------------------------
Morag Hughson
MQ Technical Education Specialist
MQGem Software Limited
Website: https://www.mqgem.com
------------------------------

Original Message

Original Message:
Sent: Wed July 13, 2022 04:51 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Morag,
The first command is giving me an error;

mqm@alphamqdev1:~ $ DEFINE LISTENER(MY.FLOATING.LSTR) TRPTYPE(TCP) IPADDR('192.0.1.58') PORT(1414) CONTROL(QM.MQT)
ksh: syntax error: unexpected '('
1|mqm@alphamqdev1:~ $

Thanks.
Regards,
Rajesh

------------------------------
RAJESH VERMA

Original Message:
Sent: Wed July 13, 2022 04:23 PM
From: Morag Hughson
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Rajesh,

As @Luc-Michel Demey said in his response before, you need to define the listener object with CONTROL(QMGR) in order for it to start automatically with the queue manager. This means the full command is as follows:-

DEFINE LISTENER(MY.FLOATING.LSTR) TRPTYPE(TCP) IPADDR('floating-ip-addr') PORT(port-number) CONTROL(QMGR)

Then the first time only, you also need to start this listener. Subsequently it will be started whenever the queue manager starts.

START LISTENER(MY.FLOATING.LSTR)

Cheers,
Morag

------------------------------
Morag Hughson
MQ Technical Education Specialist
MQGem Software Limited
Website: https://www.mqgem.com

Original Message:
Sent: Wed July 13, 2022 04:17 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Luc,
Would you please give some details as how do I define listener which automatically starts or move when RDQM starts the queue manager or move to another node.

Thank you,

Regards,
Rajesh

------------------------------
RAJESH VERMA

Original Message:
Sent: Mon July 11, 2022 11:53 AM
From: Luc-Michel Demey
Subject: RHEL Cluster -- MQ as HA cluster resource

Hello Rajesh,

A RDQM Queue Manager is just a regular Queue Manager, there is nothing special about the listener.
You need to create a listener via an MQSC command or something, and make sure it is definied with CONTROL(QMGR) so that it starts with the Queue Manager.
If the Queue Manager switches to another node, it is restarted on that node and the listener starts as usual.
Furthermore, as the IP address of the Queue Manager is a VIP moved by the RDQM infrastructure, there are no additional parameters to set in MQ Explorer. The configuration is the same as for a standalone Queue Manager.

HTH, LMD.

------------------------------
Luc-Michel Demey
DEMEY CONSULTING
lmd@demey-consulting.fr
#IBMChampion

Original Message:
Sent: Mon July 11, 2022 11:42 AM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Jamie and Morag,
As I am just system administrator with very limited knowledge about MQ, would you please guide me how can I now connect to this newly defined RDQM-Queue Manager using MQ Explorer or any other tool.
As I compared with my previous dev environment, it seems like RDQM does not run the listener program and due to I was not able to connect using MQ Explorer, the moment I run that I was able to connect to it.

But I believe that should run it alongwith the MQ when RDQM starts. Anyway I am not going to go this route to migrate any of the MQ instances (dev and prod). Now I will try to attempt migrate MQ on openshift and please be kind to let me know for any kind of procedure/steps.

Thank you very much.

Regards,
Rajesh

Original Message:
Sent: 7/6/2022 9:49:00 PM
From: RAJESH VERMA
Subject: RE: RHEL Cluster -- MQ as HA cluster resource

Hi Jamie,

I really appreciate that you took some time to answer and explain with lot of details.

Thank you very much!

With regards,

Rajesh

Original Message:
Sent: 7/6/2022 4:53:00 PM
From: Jamie Squibb
Subject: RE: RHEL Cluster -- MQ as HA cluster resource

Hi Rajesh,

> Do I really need to use three NICs for MQ -DEV -RDQM setup?

You can configure an RDQM HA group to use either one, two, or three IP addresses. An RDQM HA group has three logical interfaces (network connections) between the RDQM nodes. Two of the interfaces (primary and alternate) are used by Pacemaker to exchange heartbeats and coordinate configuration changes. The third interface is used by DRBD to replicate data between the nodes.

- If you configure a single IP address, then all three interfaces share it, which simplifies the RDQM configuration.
- If you configure two IP addresses, then one IP address is used by Pacemaker (primary and alternate) and the other IP address is used by DRBD.
- If you configure three IP addresses, then each interface has a dedicated IP address.

The number of IP addresses you choose to use depends on the environment you are using and the resilience that you require. For maximum resiliency you might choose to use three IP addresses, ensure that each IP address maps to a dedicated network card (NIC), and ensure that the network infrastructure used by the connections (switches, routers, etc.) is not shared. This configuration helps to remove single points of failure and minimise the risk of the nodes losing connectivity. However, there are other ways to achieve resiliency. For example, you might use techniques such as link aggregation to provide a single resilient network connection instead that is based on multiple physical links. This alternative approach achieves resiliency at the system/network layer, which can then be transparent to RDQM. In virtualised environments it might also be difficult to map virtual network interfaces to physical network ports, and in virtualised development environments all nodes might be hosted on the same physical server anyway.

RDQM is functionally the same for each option, so you do not need to use the same number of IP addresses in each staging environment (dev, QA, prod, etc.) if they do not require the same level of resiliency. For example, it would be common to have a greater tolerance for outages in a development environment than in production. You just need to decide whether any difference between environments is acceptable to your confidence in promoting from one environment to another.

In summary, the short answer to your question is no, you don't need to use three NICs in your development environment if you don't want to. The above longer answer provides some background for what you might want to consider when deciding if this is the correct choice for you or not.

> Do I need to put additional configuration into this newly setup DEV RDQM after importing the configuration data from AIX server using dmpmqcfg to be able to connect using MQ Explorer to queue manager in RDQM env?

I would recommend that you review MQ object definitions that refer to network connections, such as listeners and channels, to ensure they correctly reflect your RDQM setup. For example, if you are using a floating IP address you probably want to ensure that your listener(s) bind to the floating IP. This helps to ensure that applications are connecting to the floating IP instead of to a node's static IP address by mistake, which would not work after a HA failover to different node. Some specific items to check are:

- Local IP addresses used by listeners (IPADDR attribute).
- Local IP addresses used by channels (LOCLADDR attribute).
- Local IP addresses defined for cluster-receiver channels (CONNAME attribute).
- Local IP addresses defined for communication information objects (GRPADDR attribute).
- Channel authentication rules that might reference specific hostnames, IP addresses, or subnets.

You might also want to review your security configuration to determine if you need to adjust anything for user authentication and authorisation, especially if you are using local operating system userids that might need to be defined on each RDQM node.

You'll also need to review the configuration of applications and any remote queue managers that connect in to your new RDQM queue manager to ensure they no longer attempt to connect to your old AIX server.

I hope this helps.

Kind regards,
Jamie

------------------------------
Jamie Squibb
IBM MQ Development

Original Message:
Sent: Wed July 06, 2022 01:36 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Morag and Jamie,
Thank you both for being so kind. I have few more questions;
1) I have two MQ queuemanager, Dev and Prod. Do I really need to use three NICs for MQ -DEV -RDQM setup. I had setup this with one NIC on each of three Linux VM, do you think I needed three NICs for DEV MQ RDQM as well?
2) Do I need to put additional configuration into this newly setup DEV RDQM after importing the configuration data from AIX server using dmpmqcfg to be able to connect using MQ Explorer to queue manager in RDQM env.

I would definitely go with three VM - each of these with three NICs for the Production MQ RDQM environment.

I think for now I have only these questions, I will ask for more help as I continue to make progress to migrate both MQ environments to RDQM.

Thank you,
Regards,
Rajesh

------------------------------
RAJESH VERMA

Original Message:
Sent: Wed July 06, 2022 03:57 AM
From: Jamie Squibb
Subject: RHEL Cluster -- MQ as HA cluster resource

Thank you for replying Morag.

If you require any further information Rajesh, then please let us know.

Kind regards,
Jamie

------------------------------
Jamie Squibb
IBM MQ Development

Original Message:
Sent: Tue July 05, 2022 04:40 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Morag,
Thank you very much sharing the information.

Thank you,
With regards.
Rajesh Verma

------------------------------
RAJESH VERMA

Original Message:
Sent: Tue July 05, 2022 04:16 PM
From: Morag Hughson
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Rajesh,

You can use dmpmqcfg to save your MQ Configuration on the original AIX Queue Manager, and then to restore the configuration on the new RHEL Queue Manager.

If you want to copy message data as well, something like QLOAD generic unload/reload might be of interest.

Cheers,
Morag

------------------------------
Morag Hughson
MQ Technical Education Specialist
MQGem Software Limited
Website: https://www.mqgem.com

Original Message:
Sent: Tue July 05, 2022 02:07 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Jamie,
I am trying to get some help for past couple of weeks, then I see your reply to Ishara. I am working to migrate my current estate of MQ from AIX to RDQM. I have almost setup everything and looking good. Now I want to know how can I bring my current MQ configuration along with data (if possible) from AIX to this new RDQM cluster. I request you to please share some insight on this.

Thank you very much.
Regards,
Rajesh

------------------------------
RAJESH VERMA

Original Message:
Sent: Wed October 07, 2020 04:31 AM
From: Jamie Squibb
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Ishara,

I would suggest you review the Replicated Data Queue Manager (RDQM) feature that is available for RHEL 7 and RHEL 8. RDQM is a high availability and disaster recovery capability of MQ Advanced that is provided by IBM. It is built using DRBD for disk replication and Pacemaker for HA cluster management. RDQM manages the Pacemaker configuration for you.

See https://www.ibm.com/support/knowledgecenter/en/SSFKSJ_9.2.0/com.ibm.mq.con.doc/q130280_.htm

Kind regards,
Jamie

------------------------------
Jamie Squibb

Original Message:
Sent: Mon October 05, 2020 07:02 AM
From: Ishara Pathirana
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi all.

I want to get some info about how to add MQ as a cluster resource. In a Redhat Pacemaker cluster

thanks.

------------------------------
Ishara Pathirana
------------------------------

17. RE: RHEL Cluster -- MQ as HA cluster resource

Like

RAJESH VERMA

Posted Wed July 13, 2022 05:51 PM

Hi Morag,
Thank you very much it finally worked. I tested the failover and the listener also move to other node, when I failover.

Now I will try to some tests and might come back to you if I run into any other issue. I also want to experience as how to
do updates now like suspend the queue on only one node at a time and do the os updates or (not sure if we can do MQ updates
if any comes up) and document how to handle this patching part, any guidance on that is much appreciated as you are the MQ SME.

My manager also wants me to create a POC of "MQ on openshift" and then he will decided where migrate the existing MQ estate, more research huhh.

Once again thank you very much!

Regards,
Rajesh

------------------------------
RAJESH VERMA
------------------------------

Original Message

Original Message:
Sent: Wed July 13, 2022 04:52 PM
From: Morag Hughson
Subject: RHEL Cluster -- MQ as HA cluster resource

It's an MQSC command. You must issue it using the runmqsc tool (or an equivalent).

Cheers,
Morag

------------------------------
Morag Hughson
MQ Technical Education Specialist
MQGem Software Limited
Website: https://www.mqgem.com

Original Message:
Sent: Wed July 13, 2022 04:51 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Morag,
The first command is giving me an error;

mqm@alphamqdev1:~ $ DEFINE LISTENER(MY.FLOATING.LSTR) TRPTYPE(TCP) IPADDR('192.0.1.58') PORT(1414) CONTROL(QM.MQT)
ksh: syntax error: unexpected '('
1|mqm@alphamqdev1:~ $

Thanks.
Regards,
Rajesh

------------------------------
RAJESH VERMA

Original Message:
Sent: Wed July 13, 2022 04:23 PM
From: Morag Hughson
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Rajesh,

As @Luc-Michel Demey said in his response before, you need to define the listener object with CONTROL(QMGR) in order for it to start automatically with the queue manager. This means the full command is as follows:-

DEFINE LISTENER(MY.FLOATING.LSTR) TRPTYPE(TCP) IPADDR('floating-ip-addr') PORT(port-number) CONTROL(QMGR)

Then the first time only, you also need to start this listener. Subsequently it will be started whenever the queue manager starts.

START LISTENER(MY.FLOATING.LSTR)

Cheers,
Morag

------------------------------
Morag Hughson
MQ Technical Education Specialist
MQGem Software Limited
Website: https://www.mqgem.com

Original Message:
Sent: Wed July 13, 2022 04:17 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Luc,
Would you please give some details as how do I define listener which automatically starts or move when RDQM starts the queue manager or move to another node.

Thank you,

Regards,
Rajesh

------------------------------
RAJESH VERMA

Original Message:
Sent: Mon July 11, 2022 11:53 AM
From: Luc-Michel Demey
Subject: RHEL Cluster -- MQ as HA cluster resource

Hello Rajesh,

A RDQM Queue Manager is just a regular Queue Manager, there is nothing special about the listener.
You need to create a listener via an MQSC command or something, and make sure it is definied with CONTROL(QMGR) so that it starts with the Queue Manager.
If the Queue Manager switches to another node, it is restarted on that node and the listener starts as usual.
Furthermore, as the IP address of the Queue Manager is a VIP moved by the RDQM infrastructure, there are no additional parameters to set in MQ Explorer. The configuration is the same as for a standalone Queue Manager.

HTH, LMD.

------------------------------
Luc-Michel Demey
DEMEY CONSULTING
lmd@demey-consulting.fr
#IBMChampion

Original Message:
Sent: Mon July 11, 2022 11:42 AM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Jamie and Morag,
As I am just system administrator with very limited knowledge about MQ, would you please guide me how can I now connect to this newly defined RDQM-Queue Manager using MQ Explorer or any other tool.
As I compared with my previous dev environment, it seems like RDQM does not run the listener program and due to I was not able to connect using MQ Explorer, the moment I run that I was able to connect to it.

But I believe that should run it alongwith the MQ when RDQM starts. Anyway I am not going to go this route to migrate any of the MQ instances (dev and prod). Now I will try to attempt migrate MQ on openshift and please be kind to let me know for any kind of procedure/steps.

Thank you very much.

Regards,
Rajesh

Original Message:
Sent: 7/6/2022 9:49:00 PM
From: RAJESH VERMA
Subject: RE: RHEL Cluster -- MQ as HA cluster resource

Hi Jamie,

I really appreciate that you took some time to answer and explain with lot of details.

Thank you very much!

With regards,

Rajesh

Original Message:
Sent: 7/6/2022 4:53:00 PM
From: Jamie Squibb
Subject: RE: RHEL Cluster -- MQ as HA cluster resource

Hi Rajesh,

> Do I really need to use three NICs for MQ -DEV -RDQM setup?

You can configure an RDQM HA group to use either one, two, or three IP addresses. An RDQM HA group has three logical interfaces (network connections) between the RDQM nodes. Two of the interfaces (primary and alternate) are used by Pacemaker to exchange heartbeats and coordinate configuration changes. The third interface is used by DRBD to replicate data between the nodes.

- If you configure a single IP address, then all three interfaces share it, which simplifies the RDQM configuration.
- If you configure two IP addresses, then one IP address is used by Pacemaker (primary and alternate) and the other IP address is used by DRBD.
- If you configure three IP addresses, then each interface has a dedicated IP address.

The number of IP addresses you choose to use depends on the environment you are using and the resilience that you require. For maximum resiliency you might choose to use three IP addresses, ensure that each IP address maps to a dedicated network card (NIC), and ensure that the network infrastructure used by the connections (switches, routers, etc.) is not shared. This configuration helps to remove single points of failure and minimise the risk of the nodes losing connectivity. However, there are other ways to achieve resiliency. For example, you might use techniques such as link aggregation to provide a single resilient network connection instead that is based on multiple physical links. This alternative approach achieves resiliency at the system/network layer, which can then be transparent to RDQM. In virtualised environments it might also be difficult to map virtual network interfaces to physical network ports, and in virtualised development environments all nodes might be hosted on the same physical server anyway.

RDQM is functionally the same for each option, so you do not need to use the same number of IP addresses in each staging environment (dev, QA, prod, etc.) if they do not require the same level of resiliency. For example, it would be common to have a greater tolerance for outages in a development environment than in production. You just need to decide whether any difference between environments is acceptable to your confidence in promoting from one environment to another.

In summary, the short answer to your question is no, you don't need to use three NICs in your development environment if you don't want to. The above longer answer provides some background for what you might want to consider when deciding if this is the correct choice for you or not.

> Do I need to put additional configuration into this newly setup DEV RDQM after importing the configuration data from AIX server using dmpmqcfg to be able to connect using MQ Explorer to queue manager in RDQM env?

I would recommend that you review MQ object definitions that refer to network connections, such as listeners and channels, to ensure they correctly reflect your RDQM setup. For example, if you are using a floating IP address you probably want to ensure that your listener(s) bind to the floating IP. This helps to ensure that applications are connecting to the floating IP instead of to a node's static IP address by mistake, which would not work after a HA failover to different node. Some specific items to check are:

- Local IP addresses used by listeners (IPADDR attribute).
- Local IP addresses used by channels (LOCLADDR attribute).
- Local IP addresses defined for cluster-receiver channels (CONNAME attribute).
- Local IP addresses defined for communication information objects (GRPADDR attribute).
- Channel authentication rules that might reference specific hostnames, IP addresses, or subnets.

You might also want to review your security configuration to determine if you need to adjust anything for user authentication and authorisation, especially if you are using local operating system userids that might need to be defined on each RDQM node.

You'll also need to review the configuration of applications and any remote queue managers that connect in to your new RDQM queue manager to ensure they no longer attempt to connect to your old AIX server.

I hope this helps.

Kind regards,
Jamie

------------------------------
Jamie Squibb
IBM MQ Development

Original Message:
Sent: Wed July 06, 2022 01:36 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Morag and Jamie,
Thank you both for being so kind. I have few more questions;
1) I have two MQ queuemanager, Dev and Prod. Do I really need to use three NICs for MQ -DEV -RDQM setup. I had setup this with one NIC on each of three Linux VM, do you think I needed three NICs for DEV MQ RDQM as well?
2) Do I need to put additional configuration into this newly setup DEV RDQM after importing the configuration data from AIX server using dmpmqcfg to be able to connect using MQ Explorer to queue manager in RDQM env.

I would definitely go with three VM - each of these with three NICs for the Production MQ RDQM environment.

I think for now I have only these questions, I will ask for more help as I continue to make progress to migrate both MQ environments to RDQM.

Thank you,
Regards,
Rajesh

------------------------------
RAJESH VERMA

Original Message:
Sent: Wed July 06, 2022 03:57 AM
From: Jamie Squibb
Subject: RHEL Cluster -- MQ as HA cluster resource

Thank you for replying Morag.

If you require any further information Rajesh, then please let us know.

Kind regards,
Jamie

------------------------------
Jamie Squibb
IBM MQ Development

Original Message:
Sent: Tue July 05, 2022 04:40 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Morag,
Thank you very much sharing the information.

Thank you,
With regards.
Rajesh Verma

------------------------------
RAJESH VERMA

Original Message:
Sent: Tue July 05, 2022 04:16 PM
From: Morag Hughson
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Rajesh,

You can use dmpmqcfg to save your MQ Configuration on the original AIX Queue Manager, and then to restore the configuration on the new RHEL Queue Manager.

If you want to copy message data as well, something like QLOAD generic unload/reload might be of interest.

Cheers,
Morag

------------------------------
Morag Hughson
MQ Technical Education Specialist
MQGem Software Limited
Website: https://www.mqgem.com

Original Message:
Sent: Tue July 05, 2022 02:07 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Jamie,
I am trying to get some help for past couple of weeks, then I see your reply to Ishara. I am working to migrate my current estate of MQ from AIX to RDQM. I have almost setup everything and looking good. Now I want to know how can I bring my current MQ configuration along with data (if possible) from AIX to this new RDQM cluster. I request you to please share some insight on this.

Thank you very much.
Regards,
Rajesh

------------------------------
RAJESH VERMA

Original Message:
Sent: Wed October 07, 2020 04:31 AM
From: Jamie Squibb
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Ishara,

I would suggest you review the Replicated Data Queue Manager (RDQM) feature that is available for RHEL 7 and RHEL 8. RDQM is a high availability and disaster recovery capability of MQ Advanced that is provided by IBM. It is built using DRBD for disk replication and Pacemaker for HA cluster management. RDQM manages the Pacemaker configuration for you.

See https://www.ibm.com/support/knowledgecenter/en/SSFKSJ_9.2.0/com.ibm.mq.con.doc/q130280_.htm

Kind regards,
Jamie

------------------------------
Jamie Squibb

Original Message:
Sent: Mon October 05, 2020 07:02 AM
From: Ishara Pathirana
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi all.

I want to get some info about how to add MQ as a cluster resource. In a Redhat Pacemaker cluster

thanks.

------------------------------
Ishara Pathirana
------------------------------

18. RE: RHEL Cluster -- MQ as HA cluster resource

Like

RAJESH VERMA

Posted Mon July 25, 2022 06:11 PM

Hi Morag,
Would you please help me with this error which I am getting while trying to deploy queue manager on openshift;

quickstart-cp4i-ibm-mq-00/10StatefulSetSSquickstart-cp4i-ibm-mq0.1 MiB-

Jul 25, 2022, 4:49 PM

Peda-mq-lab-ibm-mq-0		0/1	0	StatefulSetSSeda-mq-lab-ibm-mq	0.1 MiB	-	Jul 25, 2022, 4:48 PM
PodPibm-mq-5cd96cbb58-wq8xd	Running	1/1	0	ReplicaSetRSibm-mq-5cd96cbb58	78.9 MiB	0.004 cores	Jul 20, 2022, 3:24 PM

Would you please help me to solve this, I am trying to create a POC to deploy MQ onto openshift.

Thank you.
Regards,
Rajesh

------------------------------
RAJESH VERMA
------------------------------

Original Message

Original Message:
Sent: Wed July 13, 2022 05:51 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Morag,
Thank you very much it finally worked. I tested the failover and the listener also move to other node, when I failover.

Now I will try to some tests and might come back to you if I run into any other issue. I also want to experience as how to
do updates now like suspend the queue on only one node at a time and do the os updates or (not sure if we can do MQ updates
if any comes up) and document how to handle this patching part, any guidance on that is much appreciated as you are the MQ SME.

My manager also wants me to create a POC of "MQ on openshift" and then he will decided where migrate the existing MQ estate, more research huhh.

Once again thank you very much!

Regards,
Rajesh

------------------------------
RAJESH VERMA

Original Message:
Sent: Wed July 13, 2022 04:52 PM
From: Morag Hughson
Subject: RHEL Cluster -- MQ as HA cluster resource

It's an MQSC command. You must issue it using the runmqsc tool (or an equivalent).

Cheers,
Morag

------------------------------
Morag Hughson
MQ Technical Education Specialist
MQGem Software Limited
Website: https://www.mqgem.com

Original Message:
Sent: Wed July 13, 2022 04:51 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Morag,
The first command is giving me an error;

mqm@alphamqdev1:~ $ DEFINE LISTENER(MY.FLOATING.LSTR) TRPTYPE(TCP) IPADDR('192.0.1.58') PORT(1414) CONTROL(QM.MQT)
ksh: syntax error: unexpected '('
1|mqm@alphamqdev1:~ $

Thanks.
Regards,
Rajesh

------------------------------
RAJESH VERMA

Original Message:
Sent: Wed July 13, 2022 04:23 PM
From: Morag Hughson
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Rajesh,

As @Luc-Michel Demey said in his response before, you need to define the listener object with CONTROL(QMGR) in order for it to start automatically with the queue manager. This means the full command is as follows:-

DEFINE LISTENER(MY.FLOATING.LSTR) TRPTYPE(TCP) IPADDR('floating-ip-addr') PORT(port-number) CONTROL(QMGR)

Then the first time only, you also need to start this listener. Subsequently it will be started whenever the queue manager starts.

START LISTENER(MY.FLOATING.LSTR)

Cheers,
Morag

------------------------------
Morag Hughson
MQ Technical Education Specialist
MQGem Software Limited
Website: https://www.mqgem.com

Original Message:
Sent: Wed July 13, 2022 04:17 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Luc,
Would you please give some details as how do I define listener which automatically starts or move when RDQM starts the queue manager or move to another node.

Thank you,

Regards,
Rajesh

------------------------------
RAJESH VERMA

Original Message:
Sent: Mon July 11, 2022 11:53 AM
From: Luc-Michel Demey
Subject: RHEL Cluster -- MQ as HA cluster resource

Hello Rajesh,

A RDQM Queue Manager is just a regular Queue Manager, there is nothing special about the listener.
You need to create a listener via an MQSC command or something, and make sure it is definied with CONTROL(QMGR) so that it starts with the Queue Manager.
If the Queue Manager switches to another node, it is restarted on that node and the listener starts as usual.
Furthermore, as the IP address of the Queue Manager is a VIP moved by the RDQM infrastructure, there are no additional parameters to set in MQ Explorer. The configuration is the same as for a standalone Queue Manager.

HTH, LMD.

------------------------------
Luc-Michel Demey
DEMEY CONSULTING
lmd@demey-consulting.fr
#IBMChampion

Original Message:
Sent: Mon July 11, 2022 11:42 AM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Jamie and Morag,
As I am just system administrator with very limited knowledge about MQ, would you please guide me how can I now connect to this newly defined RDQM-Queue Manager using MQ Explorer or any other tool.
As I compared with my previous dev environment, it seems like RDQM does not run the listener program and due to I was not able to connect using MQ Explorer, the moment I run that I was able to connect to it.

But I believe that should run it alongwith the MQ when RDQM starts. Anyway I am not going to go this route to migrate any of the MQ instances (dev and prod). Now I will try to attempt migrate MQ on openshift and please be kind to let me know for any kind of procedure/steps.

Thank you very much.

Regards,
Rajesh

Original Message:
Sent: 7/6/2022 9:49:00 PM
From: RAJESH VERMA
Subject: RE: RHEL Cluster -- MQ as HA cluster resource

Hi Jamie,

I really appreciate that you took some time to answer and explain with lot of details.

Thank you very much!

With regards,

Rajesh

Original Message:
Sent: 7/6/2022 4:53:00 PM
From: Jamie Squibb
Subject: RE: RHEL Cluster -- MQ as HA cluster resource

Hi Rajesh,

> Do I really need to use three NICs for MQ -DEV -RDQM setup?

You can configure an RDQM HA group to use either one, two, or three IP addresses. An RDQM HA group has three logical interfaces (network connections) between the RDQM nodes. Two of the interfaces (primary and alternate) are used by Pacemaker to exchange heartbeats and coordinate configuration changes. The third interface is used by DRBD to replicate data between the nodes.

- If you configure a single IP address, then all three interfaces share it, which simplifies the RDQM configuration.
- If you configure two IP addresses, then one IP address is used by Pacemaker (primary and alternate) and the other IP address is used by DRBD.
- If you configure three IP addresses, then each interface has a dedicated IP address.

The number of IP addresses you choose to use depends on the environment you are using and the resilience that you require. For maximum resiliency you might choose to use three IP addresses, ensure that each IP address maps to a dedicated network card (NIC), and ensure that the network infrastructure used by the connections (switches, routers, etc.) is not shared. This configuration helps to remove single points of failure and minimise the risk of the nodes losing connectivity. However, there are other ways to achieve resiliency. For example, you might use techniques such as link aggregation to provide a single resilient network connection instead that is based on multiple physical links. This alternative approach achieves resiliency at the system/network layer, which can then be transparent to RDQM. In virtualised environments it might also be difficult to map virtual network interfaces to physical network ports, and in virtualised development environments all nodes might be hosted on the same physical server anyway.

RDQM is functionally the same for each option, so you do not need to use the same number of IP addresses in each staging environment (dev, QA, prod, etc.) if they do not require the same level of resiliency. For example, it would be common to have a greater tolerance for outages in a development environment than in production. You just need to decide whether any difference between environments is acceptable to your confidence in promoting from one environment to another.

In summary, the short answer to your question is no, you don't need to use three NICs in your development environment if you don't want to. The above longer answer provides some background for what you might want to consider when deciding if this is the correct choice for you or not.

> Do I need to put additional configuration into this newly setup DEV RDQM after importing the configuration data from AIX server using dmpmqcfg to be able to connect using MQ Explorer to queue manager in RDQM env?

I would recommend that you review MQ object definitions that refer to network connections, such as listeners and channels, to ensure they correctly reflect your RDQM setup. For example, if you are using a floating IP address you probably want to ensure that your listener(s) bind to the floating IP. This helps to ensure that applications are connecting to the floating IP instead of to a node's static IP address by mistake, which would not work after a HA failover to different node. Some specific items to check are:

- Local IP addresses used by listeners (IPADDR attribute).
- Local IP addresses used by channels (LOCLADDR attribute).
- Local IP addresses defined for cluster-receiver channels (CONNAME attribute).
- Local IP addresses defined for communication information objects (GRPADDR attribute).
- Channel authentication rules that might reference specific hostnames, IP addresses, or subnets.

You might also want to review your security configuration to determine if you need to adjust anything for user authentication and authorisation, especially if you are using local operating system userids that might need to be defined on each RDQM node.

You'll also need to review the configuration of applications and any remote queue managers that connect in to your new RDQM queue manager to ensure they no longer attempt to connect to your old AIX server.

I hope this helps.

Kind regards,
Jamie

------------------------------
Jamie Squibb
IBM MQ Development

Original Message:
Sent: Wed July 06, 2022 01:36 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Morag and Jamie,
Thank you both for being so kind. I have few more questions;
1) I have two MQ queuemanager, Dev and Prod. Do I really need to use three NICs for MQ -DEV -RDQM setup. I had setup this with one NIC on each of three Linux VM, do you think I needed three NICs for DEV MQ RDQM as well?
2) Do I need to put additional configuration into this newly setup DEV RDQM after importing the configuration data from AIX server using dmpmqcfg to be able to connect using MQ Explorer to queue manager in RDQM env.

I would definitely go with three VM - each of these with three NICs for the Production MQ RDQM environment.

I think for now I have only these questions, I will ask for more help as I continue to make progress to migrate both MQ environments to RDQM.

Thank you,
Regards,
Rajesh

------------------------------
RAJESH VERMA

Original Message:
Sent: Wed July 06, 2022 03:57 AM
From: Jamie Squibb
Subject: RHEL Cluster -- MQ as HA cluster resource

Thank you for replying Morag.

If you require any further information Rajesh, then please let us know.

Kind regards,
Jamie

------------------------------
Jamie Squibb
IBM MQ Development

Original Message:
Sent: Tue July 05, 2022 04:40 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Morag,
Thank you very much sharing the information.

Thank you,
With regards.
Rajesh Verma

------------------------------
RAJESH VERMA

Original Message:
Sent: Tue July 05, 2022 04:16 PM
From: Morag Hughson
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Rajesh,

You can use dmpmqcfg to save your MQ Configuration on the original AIX Queue Manager, and then to restore the configuration on the new RHEL Queue Manager.

If you want to copy message data as well, something like QLOAD generic unload/reload might be of interest.

Cheers,
Morag

------------------------------
Morag Hughson
MQ Technical Education Specialist
MQGem Software Limited
Website: https://www.mqgem.com

Original Message:
Sent: Tue July 05, 2022 02:07 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Jamie,
I am trying to get some help for past couple of weeks, then I see your reply to Ishara. I am working to migrate my current estate of MQ from AIX to RDQM. I have almost setup everything and looking good. Now I want to know how can I bring my current MQ configuration along with data (if possible) from AIX to this new RDQM cluster. I request you to please share some insight on this.

Thank you very much.
Regards,
Rajesh

------------------------------
RAJESH VERMA

Original Message:
Sent: Wed October 07, 2020 04:31 AM
From: Jamie Squibb
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Ishara,

I would suggest you review the Replicated Data Queue Manager (RDQM) feature that is available for RHEL 7 and RHEL 8. RDQM is a high availability and disaster recovery capability of MQ Advanced that is provided by IBM. It is built using DRBD for disk replication and Pacemaker for HA cluster management. RDQM manages the Pacemaker configuration for you.

See https://www.ibm.com/support/knowledgecenter/en/SSFKSJ_9.2.0/com.ibm.mq.con.doc/q130280_.htm

Kind regards,
Jamie

------------------------------
Jamie Squibb

Original Message:
Sent: Mon October 05, 2020 07:02 AM
From: Ishara Pathirana
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi all.

I want to get some info about how to add MQ as a cluster resource. In a Redhat Pacemaker cluster

thanks.

------------------------------
Ishara Pathirana
------------------------------

19. RE: RHEL Cluster -- MQ as HA cluster resource

Like

Morag Hughson

IBM Champion

Posted Mon July 25, 2022 06:15 PM

Hi Rajesh,

Could you please check that you have correctly conveyed the error you are seeing. Your post has a weird mixture of hyperlinks and images. I am unclear what the error you are seeing is.

Also, if this is nothing to do with a RHEL cluster, perhaps it would be better to start a new thread with an appropriate title, and let everyone read it and provide assistance. I am not an OpenShift expert.

Cheers,
Morag

------------------------------
Morag Hughson
MQ Technical Education Specialist
MQGem Software Limited
Website: https://www.mqgem.com
------------------------------

Original Message

Original Message:
Sent: Mon July 25, 2022 06:10 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Morag,
Would you please help me with this error which I am getting while trying to deploy queue manager on openshift;

quickstart-cp4i-ibm-mq-00/10StatefulSetSSquickstart-cp4i-ibm-mq0.1 MiB-

Jul 25, 2022, 4:49 PM

Peda-mq-lab-ibm-mq-0		0/1	0	StatefulSetSSeda-mq-lab-ibm-mq	0.1 MiB	-	Jul 25, 2022, 4:48 PM
PodPibm-mq-5cd96cbb58-wq8xd	Running	1/1	0	ReplicaSetRSibm-mq-5cd96cbb58	78.9 MiB	0.004 cores	Jul 20, 2022, 3:24 PM

Would you please help me to solve this, I am trying to create a POC to deploy MQ onto openshift.

Thank you.
Regards,
Rajesh

------------------------------
RAJESH VERMA

Original Message:
Sent: Wed July 13, 2022 05:51 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Morag,
Thank you very much it finally worked. I tested the failover and the listener also move to other node, when I failover.

Now I will try to some tests and might come back to you if I run into any other issue. I also want to experience as how to
do updates now like suspend the queue on only one node at a time and do the os updates or (not sure if we can do MQ updates
if any comes up) and document how to handle this patching part, any guidance on that is much appreciated as you are the MQ SME.

My manager also wants me to create a POC of "MQ on openshift" and then he will decided where migrate the existing MQ estate, more research huhh.

Once again thank you very much!

Regards,
Rajesh

------------------------------
RAJESH VERMA

Original Message:
Sent: Wed July 13, 2022 04:52 PM
From: Morag Hughson
Subject: RHEL Cluster -- MQ as HA cluster resource

It's an MQSC command. You must issue it using the runmqsc tool (or an equivalent).

Cheers,
Morag

------------------------------
Morag Hughson
MQ Technical Education Specialist
MQGem Software Limited
Website: https://www.mqgem.com

Original Message:
Sent: Wed July 13, 2022 04:51 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Morag,
The first command is giving me an error;

mqm@alphamqdev1:~ $ DEFINE LISTENER(MY.FLOATING.LSTR) TRPTYPE(TCP) IPADDR('192.0.1.58') PORT(1414) CONTROL(QM.MQT)
ksh: syntax error: unexpected '('
1|mqm@alphamqdev1:~ $

Thanks.
Regards,
Rajesh

------------------------------
RAJESH VERMA

Original Message:
Sent: Wed July 13, 2022 04:23 PM
From: Morag Hughson
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Rajesh,

As @Luc-Michel Demey said in his response before, you need to define the listener object with CONTROL(QMGR) in order for it to start automatically with the queue manager. This means the full command is as follows:-

DEFINE LISTENER(MY.FLOATING.LSTR) TRPTYPE(TCP) IPADDR('floating-ip-addr') PORT(port-number) CONTROL(QMGR)

Then the first time only, you also need to start this listener. Subsequently it will be started whenever the queue manager starts.

START LISTENER(MY.FLOATING.LSTR)

Cheers,
Morag

------------------------------
Morag Hughson
MQ Technical Education Specialist
MQGem Software Limited
Website: https://www.mqgem.com

Original Message:
Sent: Wed July 13, 2022 04:17 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Luc,
Would you please give some details as how do I define listener which automatically starts or move when RDQM starts the queue manager or move to another node.

Thank you,

Regards,
Rajesh

------------------------------
RAJESH VERMA

Original Message:
Sent: Mon July 11, 2022 11:53 AM
From: Luc-Michel Demey
Subject: RHEL Cluster -- MQ as HA cluster resource

Hello Rajesh,

A RDQM Queue Manager is just a regular Queue Manager, there is nothing special about the listener.
You need to create a listener via an MQSC command or something, and make sure it is definied with CONTROL(QMGR) so that it starts with the Queue Manager.
If the Queue Manager switches to another node, it is restarted on that node and the listener starts as usual.
Furthermore, as the IP address of the Queue Manager is a VIP moved by the RDQM infrastructure, there are no additional parameters to set in MQ Explorer. The configuration is the same as for a standalone Queue Manager.

HTH, LMD.

------------------------------
Luc-Michel Demey
DEMEY CONSULTING
lmd@demey-consulting.fr
#IBMChampion

Original Message:
Sent: Mon July 11, 2022 11:42 AM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Jamie and Morag,
As I am just system administrator with very limited knowledge about MQ, would you please guide me how can I now connect to this newly defined RDQM-Queue Manager using MQ Explorer or any other tool.
As I compared with my previous dev environment, it seems like RDQM does not run the listener program and due to I was not able to connect using MQ Explorer, the moment I run that I was able to connect to it.

But I believe that should run it alongwith the MQ when RDQM starts. Anyway I am not going to go this route to migrate any of the MQ instances (dev and prod). Now I will try to attempt migrate MQ on openshift and please be kind to let me know for any kind of procedure/steps.

Thank you very much.

Regards,
Rajesh

Original Message:
Sent: 7/6/2022 9:49:00 PM
From: RAJESH VERMA
Subject: RE: RHEL Cluster -- MQ as HA cluster resource

Hi Jamie,

I really appreciate that you took some time to answer and explain with lot of details.

Thank you very much!

With regards,

Rajesh

Original Message:
Sent: 7/6/2022 4:53:00 PM
From: Jamie Squibb
Subject: RE: RHEL Cluster -- MQ as HA cluster resource

Hi Rajesh,

> Do I really need to use three NICs for MQ -DEV -RDQM setup?

You can configure an RDQM HA group to use either one, two, or three IP addresses. An RDQM HA group has three logical interfaces (network connections) between the RDQM nodes. Two of the interfaces (primary and alternate) are used by Pacemaker to exchange heartbeats and coordinate configuration changes. The third interface is used by DRBD to replicate data between the nodes.

- If you configure a single IP address, then all three interfaces share it, which simplifies the RDQM configuration.
- If you configure two IP addresses, then one IP address is used by Pacemaker (primary and alternate) and the other IP address is used by DRBD.
- If you configure three IP addresses, then each interface has a dedicated IP address.

The number of IP addresses you choose to use depends on the environment you are using and the resilience that you require. For maximum resiliency you might choose to use three IP addresses, ensure that each IP address maps to a dedicated network card (NIC), and ensure that the network infrastructure used by the connections (switches, routers, etc.) is not shared. This configuration helps to remove single points of failure and minimise the risk of the nodes losing connectivity. However, there are other ways to achieve resiliency. For example, you might use techniques such as link aggregation to provide a single resilient network connection instead that is based on multiple physical links. This alternative approach achieves resiliency at the system/network layer, which can then be transparent to RDQM. In virtualised environments it might also be difficult to map virtual network interfaces to physical network ports, and in virtualised development environments all nodes might be hosted on the same physical server anyway.

RDQM is functionally the same for each option, so you do not need to use the same number of IP addresses in each staging environment (dev, QA, prod, etc.) if they do not require the same level of resiliency. For example, it would be common to have a greater tolerance for outages in a development environment than in production. You just need to decide whether any difference between environments is acceptable to your confidence in promoting from one environment to another.

In summary, the short answer to your question is no, you don't need to use three NICs in your development environment if you don't want to. The above longer answer provides some background for what you might want to consider when deciding if this is the correct choice for you or not.

> Do I need to put additional configuration into this newly setup DEV RDQM after importing the configuration data from AIX server using dmpmqcfg to be able to connect using MQ Explorer to queue manager in RDQM env?

I would recommend that you review MQ object definitions that refer to network connections, such as listeners and channels, to ensure they correctly reflect your RDQM setup. For example, if you are using a floating IP address you probably want to ensure that your listener(s) bind to the floating IP. This helps to ensure that applications are connecting to the floating IP instead of to a node's static IP address by mistake, which would not work after a HA failover to different node. Some specific items to check are:

- Local IP addresses used by listeners (IPADDR attribute).
- Local IP addresses used by channels (LOCLADDR attribute).
- Local IP addresses defined for cluster-receiver channels (CONNAME attribute).
- Local IP addresses defined for communication information objects (GRPADDR attribute).
- Channel authentication rules that might reference specific hostnames, IP addresses, or subnets.

You might also want to review your security configuration to determine if you need to adjust anything for user authentication and authorisation, especially if you are using local operating system userids that might need to be defined on each RDQM node.

You'll also need to review the configuration of applications and any remote queue managers that connect in to your new RDQM queue manager to ensure they no longer attempt to connect to your old AIX server.

I hope this helps.

Kind regards,
Jamie

------------------------------
Jamie Squibb
IBM MQ Development

Original Message:
Sent: Wed July 06, 2022 01:36 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Morag and Jamie,
Thank you both for being so kind. I have few more questions;
1) I have two MQ queuemanager, Dev and Prod. Do I really need to use three NICs for MQ -DEV -RDQM setup. I had setup this with one NIC on each of three Linux VM, do you think I needed three NICs for DEV MQ RDQM as well?
2) Do I need to put additional configuration into this newly setup DEV RDQM after importing the configuration data from AIX server using dmpmqcfg to be able to connect using MQ Explorer to queue manager in RDQM env.

I would definitely go with three VM - each of these with three NICs for the Production MQ RDQM environment.

I think for now I have only these questions, I will ask for more help as I continue to make progress to migrate both MQ environments to RDQM.

Thank you,
Regards,
Rajesh

------------------------------
RAJESH VERMA

Original Message:
Sent: Wed July 06, 2022 03:57 AM
From: Jamie Squibb
Subject: RHEL Cluster -- MQ as HA cluster resource

Thank you for replying Morag.

If you require any further information Rajesh, then please let us know.

Kind regards,
Jamie

------------------------------
Jamie Squibb
IBM MQ Development

Original Message:
Sent: Tue July 05, 2022 04:40 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Morag,
Thank you very much sharing the information.

Thank you,
With regards.
Rajesh Verma

------------------------------
RAJESH VERMA

Original Message:
Sent: Tue July 05, 2022 04:16 PM
From: Morag Hughson
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Rajesh,

You can use dmpmqcfg to save your MQ Configuration on the original AIX Queue Manager, and then to restore the configuration on the new RHEL Queue Manager.

If you want to copy message data as well, something like QLOAD generic unload/reload might be of interest.

Cheers,
Morag

------------------------------
Morag Hughson
MQ Technical Education Specialist
MQGem Software Limited
Website: https://www.mqgem.com

Original Message:
Sent: Tue July 05, 2022 02:07 PM
From: RAJESH VERMA
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Jamie,
I am trying to get some help for past couple of weeks, then I see your reply to Ishara. I am working to migrate my current estate of MQ from AIX to RDQM. I have almost setup everything and looking good. Now I want to know how can I bring my current MQ configuration along with data (if possible) from AIX to this new RDQM cluster. I request you to please share some insight on this.

Thank you very much.
Regards,
Rajesh

------------------------------
RAJESH VERMA

Original Message:
Sent: Wed October 07, 2020 04:31 AM
From: Jamie Squibb
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi Ishara,

I would suggest you review the Replicated Data Queue Manager (RDQM) feature that is available for RHEL 7 and RHEL 8. RDQM is a high availability and disaster recovery capability of MQ Advanced that is provided by IBM. It is built using DRBD for disk replication and Pacemaker for HA cluster management. RDQM manages the Pacemaker configuration for you.

See https://www.ibm.com/support/knowledgecenter/en/SSFKSJ_9.2.0/com.ibm.mq.con.doc/q130280_.htm

Kind regards,
Jamie

------------------------------
Jamie Squibb

Original Message:
Sent: Mon October 05, 2020 07:02 AM
From: Ishara Pathirana
Subject: RHEL Cluster -- MQ as HA cluster resource

Hi all.

I want to get some info about how to add MQ as a cluster resource. In a Redhat Pacemaker cluster

thanks.

------------------------------
Ishara Pathirana
------------------------------

20. RE: RHEL Cluster -- MQ as HA cluster resource

Like

RAJESH VERMA

Posted Mon April 03, 2023 03:32 PM

Hi Morag,
The "MQ deployment in Openshift" hit the dead-end due to the limitation in the application. I am going to work back on the RDQM Cluster, I want to ask you couple questions;

1) Will ONE CPU (Intel(R) Xeon(R) Gold 6234 CPU @ 3.30GHz with 8GB RAM) be sufficient to create the RDQM cluster on each of the three nodes especially in Development environment?
2) How big the queue size should I give while defining queuemanager (like crtmqm -sxs -fs 4096M QM.MQT), can I use 16G size in there.

Thank you very much!!!
Regards,

Rajesh

------------------------------
RAJESH VERMA
------------------------------

Original Message

MQ

MQ