| Author |
Message
|
| fjb_saper |
 Posted: Tue Sep 01, 2020 10:39 am Post subject: |
 |
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
@exerk
Why 3 replication sets (9 IPS) instead of one set (3 IPS) with rotating primary secondary and replication ?? 
_________________
MQ & Broker admin |
|
| Back to top |
|
 |
| exerk |
| Posted: Tue Sep 01, 2020 11:10 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
This was a Proof-of-Concept I set up over two years ago so I followed the KC faithfully:
| Quote: |
| The rdqm.ini file gives the IP addresses for all of the nodes in the Pacemaker cluster. You can specify that the Pacemaker cluster uses one, two, or three IP addresses. The interface that is used for synchronous data replication is named the 'replication interface'. The interface must have sufficient bandwidth to support replication requirements given the expected workload of all the RDQMs running in the HA Group. The primary and secondary interfaces are used for the Pacemaker to monitor the system, but Pacemaker can use the replication interface for this purpose, if required. |
I'm not sure what you mean by "...rotating primary secondary and replication...", sorry.
_________________
It's puzzling, I don't think I've ever seen anything quite like this before...and it's hard to soar like an eagle when you're surrounded by turkeys. |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Wed Sep 02, 2020 4:57 am Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
| exerk wrote: |
I'm not sure what you mean by "...rotating primary secondary and replication...", sorry. |
Meaning that you have 3 IPS as primary, secondary and replication
Meaning that you define 3 qmgrs QM1 QM2 QM3
and preffered node is
QM1 primary
QM2 secondary
QM3 replication
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| crashdog |
| Posted: Wed Sep 02, 2020 5:04 am Post subject: |
|
|
Voyager
Joined: 02 Apr 2017
Posts: 77
|
That's not really how it works... You have 3 nodes (VMs). Basically you only need two network cards per node, one for the floating ip and one for the replication.
The primary and alternate nics are only for the heartbeat.
The number of queue managers per node is not limited. (it used to be limited to 10 and IBM's RDQM performance tests use up to 10 qmgrs).
Cheers,
Gerhard
_________________
You win again gravity ! |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Wed Sep 02, 2020 5:42 am Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
| crashdog wrote: |
That's not really how it works... You have 3 nodes (VMs). Basically you only need two network cards per node, one for the floating ip and one for the replication.
The primary and alternate nics are only for the heartbeat.
The number of queue managers per node is not limited. (it used to be limited to 10 and IBM's RDQM performance tests use up to 10 qmgrs).
Cheers,
Gerhard |
Thanks for enlightening me. I naively thought that one network card could handle both the node's ip and the floating ip...
I understand the use of an alternate card for the heartbeat.
So if I understand this right Exerk is using 3 network cards
1 for the primary traffic
1 for the heartbeat traffic
1 for the replication traffic.
Please correct me if I misunderstood this set up
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| crashdog |
| Posted: Wed Sep 02, 2020 5:56 am Post subject: |
|
|
Voyager
Joined: 02 Apr 2017
Posts: 77
|
you got it 
_________________
You win again gravity ! |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Wed Sep 02, 2020 12:06 pm Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
| crashdog wrote: |
| you got it |
I was just a little bit surprised because usually each of those cards would be in a different subnet...
1 subnet for primary traffic
1 subnet for heartbeat
1 subnet for replications
And Exerk had them all on the same subnet... not that there's anything wrong with that...
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| exerk |
| Posted: Thu Sep 03, 2020 10:04 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| fjb_saper wrote: |
| ...And Exerk had them all on the same subnet... not that there's anything wrong with that... |
It was a test installation Proof-of-Concept in a VMWare environment - not worth the hassle of creating a 'full' network infrastructure just to test out the mechanics of RDQM 
_________________
It's puzzling, I don't think I've ever seen anything quite like this before...and it's hard to soar like an eagle when you're surrounded by turkeys. |
|
| Back to top |
|
|
| jinyuximu |
| Posted: Tue Jun 22, 2021 4:06 am Post subject: |
|
|
Guest
|
The problem I encountered is the same as yours. I know how to do it now.
My environment is: Centos8.3 and MQ 9.2.2CD. The firewall is turned off.
I vi /etc/selinux/config and SELINUX=disabled. Then reboot 3 machines. Execute crtmqm -sx RDQM_HA1. rdqmstatus -m RDQM_HA1 is fine.
| crashdog wrote: |
| Quote: |
| It works, trust me... |
Not that I would ever doubt your trustworthiness but, nope same issue with REHL 7.8 and MQ 9.2.0.0 as CentOS 7.8 and MQ 9.2.0.0.
Just wondering what I could possibly be doing wrong... I've checked just about everything from drbd over corosync to pacemaker. I went over the MQ installation guide time and time again...
I guess I will try the RHEL / MQ9.2.0.0 installation on my employers system, if the issue occurs there too I guess I will just have to open a PMR.
EDIT: Is it possible to open PMRs or have any official IBM support as a "private" person ? means I don't have an MQ license. Just using the developer editions for educational purposes.
Cheers,
Gerhard |
|
|
| Back to top |
|
|
|
|
