I have both MQ queue managers and MQ appliances in MQ clusters, with one instance in standby and the other active.  I am using APM 8.1.4 MQ agents to monitor each instance.  How can I determine when one MQ instance has gone inactive and failed over to the other instance?  Is there a particular MQ event that I can look for?

I don't know of any event messages that MQ might generate during/after failover but can't you get Tivoli to periodically issue DSPMQ on each instance ?

The Queue Manager Active event message was enhanced when multi-instance queue manager's were created to include the hostname that the queue manager is running on. This allows you to see which instance of the queue manager has just become active.

Aha ! Thank you Morag. I didn't know this.

Thanks, Morag.  Unfortunately that does not appear to be a valid reason code in APM.  It might be that they're using a different name in the situation definition but it will take a little while for me to figure out the enumeration.

I might need to open a PMR to get the detailed information I need.

It surprises me that IBM APM 8.1.4 doesn't support Queue Manager Active events. That event has been around since day 1 of MQ. What events do you currently monitor?

Queue Manager Active is certainly mentioned in the docs for APM when you search for both, but I'm not familiar with the way APM configures monitoring for events.

On ITM V6 the queue managers status is "enumerated" because I believe MQ holds the real value of that attribute as a numeric value. 1, 2, 3 etc to represent the status. ITM enumerates the values to "Running", "Active", "Inactive" but APM does not. It requires that the user compare on the numeric value. I have an RFE opened for this. I don't know when or if it will be addressed.

Morag - the APM agent DOES support "Queue Manager Active".  But it's the reason code that's missing.  There are tons of valid reason codes for that event, but none of them are for the "standby became active" reason code.

Steve - the enumeration I can deal with; I can find the actual event values.  What is the KMK agent?  I was not aware that there was support for the MQ appliance in ITM 6, and there's not an MK agent listed in the product codes.  The only agent install option in the APM Advanced agent package is "Websphere MQ".

Joe the KMK agent is the formal name for the MQ Appliance Monitoring Agent.
It's very new and is undergoing productization. We are working to get it on
Passport etc. The agent is designed to monitor the appliance hardware and
queue manager "status" on the appliance(s). It's built on the Agent Builder
REST API provider. I worked with Hursley to build this agent and it is
fully supported now and is running in production at US Customs and Border
Protection.

At this time the agent does not provide MQSC data (channel info, queue
manager stats etc) like the Linux / Windows based MQ agent so they have to
work in tandem to get all of the info. You have to create the CCDT on the
appliance and download it to where the soft MQ agent is running. This is
temporary as we are working to put the MQSC data in the KMK agent. If you
are looking for EVENT data you either have to pull that with the standard
MQ agent.

KMK works with ITM V6 and APM V8.

I mentioned the queue manager status information as I thought that's what
you were looking for.

Here's a screen shot for ITM V6 using a custom set of workspaces.




Steve Coleman
Hybrid Cloud ITSM Technical Sales Specialist Federal
US Army Intelligence And Security Command (Retired)
stcolema@us.ibm.com
(717) 466-6257 (Home)
(717) 208-0857 (Mobile)




From:"Joe Howell, Dick's Sporting Goods, Inc."
<wsmqfam-ws@lists.imwuc.org>
To:WSMQFam-ws@lists.imwuc.org
Date:04/02/2018 09:20 AM
Subject:[WSMQFam-ws] - RE: How can I determine when one MQ instance has
gone inactive and failed over to the other instance?



Morag - the APM agent DOES support "Queue Manager Active".  But it's the
reason code that's missing.  There are tons of valid reason codes for that
event, but none of them are for the "standby became active" reason code.


Steve - the enumeration I can deal with; I can find the actual event
values.  What is the KMK agent?  I was not aware that there was support for
the MQ appliance in ITM 6, and there's not an MK agent listed in the
product codes.  The only agent install option in the APM Advanced agent
package is "Websphere MQ".




Site Links: View post online View mailing list online Start new thread
via email Unsubscribe from this mailing list Manage your subscription


This email has been sent to: stcolema@us.ibm.com

Outstanding - that's probably exactly what I need.  I'll keep an eye out for it.  Thanks.

If you want it early I can get it to you and we can configure it.


Steve Coleman
Hybrid Cloud ITSM Technical Sales Specialist Federal
US Army Intelligence And Security Command (Retired)
stcolema@us.ibm.com
(717) 466-6257 (Home)
(717) 208-0857 (Mobile)




From:"Joe Howell, Dick's Sporting Goods, Inc."
<wsmqfam-ws@lists.imwuc.org>
To:WSMQFam-ws@lists.imwuc.org
Date:04/02/2018 10:34 AM
Subject:[WSMQFam-ws] - RE: How can I determine when one MQ instance has
gone inactive and failed over to the other instance?



Outstanding - that's probably exactly what I need.  I'll keep an eye out
for it.  Thanks.




Site Links: View post online View mailing list online Start new thread
via email Unsubscribe from this mailing list Manage your subscription


This email has been sent to: stcolema@us.ibm.com

I appreciate the offer but I'm going to have to decline.  I'm pretty much a one-man operation and need to keep things as simple as I can.  Plus my management is allergic to "early" code.  :)

I can afford to wait for the product to be released in the APM bundle.

Ok. No worries. It's not early code as there is no code. The Agent Builder
provides multiple types of target data sources for monitoring. WBEM / WMI,
Windows Perfmon, JDBC queries, HTTP/S queries, REST API queries, log
monitoring ..to name a few. The KMK agent is built on the REST API query.
You give it the URI endpoint and the REST query and it returns the results.
Fully Supported IBM Raleigh. It's going to be delivered on with ITM V6
agent packages and APM V8 agent packages. The administrative stuff takes a
while.

It you guys wanna' check it out let me know. I built it for Hursley and
they send all of the monitoring requests / requirements to me.

If you haven't used the Agent Builder...check it out. It's free if you are
license for ITM or APM.

Later Joe.


Steve Coleman
Hybrid Cloud ITSM Technical Sales Specialist Federal
US Army Intelligence And Security Command (Retired)
stcolema@us.ibm.com
(717) 466-6257 (Home)
(717) 208-0857 (Mobile)




From:"Joe Howell, Dick's Sporting Goods, Inc."
<wsmqfam-ws@lists.imwuc.org>
To:WSMQFam-ws@lists.imwuc.org
Date:04/02/2018 11:04 AM
Subject:[WSMQFam-ws] - RE: How can I determine when one MQ instance has
gone inactive and failed over to the other instance?



I appreciate the offer but I'm going to have to decline.  I'm pretty much a
one-man operation and need to keep things as simple as I can.  Plus my
management is allergic to "early" code.  :)


I can afford to wait for the product to be released in the APM bundle.




Site Links: View post online View mailing list online Start new thread
via email Unsubscribe from this mailing list Manage your subscription


This email has been sent to: stcolema@us.ibm.com

We help our customers design, and manage very small, to very large MQ environments (and actively manage and monitor as well as an MSP)...  We would normally suggest this for most shops.  In our case, we use Avada's IR 360 products, in conjunction with others to not just provide alerting, but also, again used in conjunction with our services, to do trend analysis, and predictive analytics.

Hi Joe. Steve Coleman here. The KMK MQ Appliance Agent monitors both the primary and standby queue managers. On the appliance if the primary queue manager is not in a "HARole" of "Primary" and a "Status" of "Running" then there is an issue with the primary queue manager. If the standby queue manager is not in a "Role" of "Secondary" and a "Status" of "Running elsewhere" then there's a problem with the standby queue manager.

I mention the above because you could have a condition where the primary is no longer acting as the primary yet the standby hasn't assumed responsibility.

If you see the standby queue manager in a "Role" of "Primary" and a "Status" of "Running" then a failover has occurred.  The standby should always be in "Secondary" and "Running elsewhere" mode.

Situations for ITM V6 and threholds for APM V8 should allow for the monitoring of the above conditions. I personally create situations and thresholds that provide an "Ok" condition as well.

Let me know if I can help.