ASG
IBM
Zystems
Cressida
Icon
Netflexity
 
  MQSeries.net
Search  Search       Tech Exchange      Education      Certifications      Library      Info Center      SupportPacs      LinkedIn  Search  Search                                                                   FAQ  FAQ   Usergroups  Usergroups
 
Register  ::  Log in Log in to check your private messages
 
RSS Feed - WebSphere MQ Support RSS Feed - Message Broker Support

MQSeries.net Forum Index » Clustering » Not able to add z/OS queue manager to cluster

Post new topic  Reply to topic Goto page 1, 2  Next
 Not able to add z/OS queue manager to cluster « View previous topic :: View next topic » 
Author Message
Prithviraj
PostPosted: Thu May 29, 2014 4:38 am    Post subject: Not able to add z/OS queue manager to cluster Reply with quote

Apprentice

Joined: 29 May 2014
Posts: 30

Dear all,

I have created a cluster (NCLUSTER) b/w queue managers from windows and z/os. Cluster definitions are as follows

Windows queue manager(winodws12) definitions :

ALTER QMGR REPOS(NCLUSTER)

DEFINE CHANNEL(TO.WINDOWS12) CHLTYPE(CLUSRCVR) TRPTYPE(TCP) CONNAME('172.16.111.222(1414)') CLUSTER(NCLUSTER)

DEFINE CHANNEL(FROM.WINDOWS12) CHLTYPE(CLUSSDR) TRPTYPE(TCP) CONNAME('172.16.100.200(1400)') CLUSTER(NCLUSTER)

z/OS queue manager(ZOS1) defenitions:

DEFINE CHANNEL(TO.WINDOWS12) CHLTYPE(CLUSSDR) TRPTYPE(TCP) CONNAME(172.16.111.222(1414)) CLUSTER(NCLUSTER)

DEFINE CHANNEL(FROM.WINDOWS12) CHLTYPE(CLUSRCVR) TRPTYPE(TCP) CONNAME('172.16.100.200(1400)') CLUSTER(NCLUSTER)

Now channel status in windows is

display chstatus(*)
17 : display chstatus(*)
AMQ8417: Display Channel Status details.
CHANNEL(FROM.WINDOWS12) CHLTYPE(CLUSSDR)
CONNAME(172.16.100.200(1400)) CURRENT
RQMNAME(ZOS1) STATUS(RUNNING)
SUBSTATE(MQGET) XMITQ(SYSTEM.CLUSTER.TRANSMIT.QUEUE)
AMQ8417: Display Channel Status details.
CHANNEL(TO.WINDOWS12) CHLTYPE(CLUSRCVR)
CONNAME(172.16.100.200) CURRENT
RQMNAME(ZOS1) STATUS(RUNNING)
SUBSTATE(RECEIVE)

Now when we give dis clusqmgr(*) in cmd in windows we get

dis clusqmgr(*)
18 : dis clusqmgr(*)
AMQ8441: Display Cluster Queue Manager details.
CLUSQMGR(SYSTEM.TEMPQMGR.172.16.100.200(1400))
CHANNEL(FROM.WINDOWS12) CLUSTER(NCLUSTER)
AMQ8441: Display Cluster Queue Manager details.
CLUSQMGR(WINDOWS12) CHANNEL(TO.WINDOWS12)
CLUSTER(NCLUSTER)

From MQ Explorer I got the message "Severity Description Object name Category
Error Cluster NCLUSTER is unable to contact qmgr believed to be at '172.16.100.200(1400)' (using channel FROM.WINDOWS12) WINDOWS12 Queue Manager / Clusters"


My channel definitions are correct.

Now my problem is queue manager from z/OS is not coming under cluster .

Now if I add 2 windows queue managers as full repositories and z/OS queue manager as partial repositories still it is not working(earlier it worked fine).

When I add z/OS queue manager as partial repos the clssdr channels are running but cluster-receiver channel(z/OS queue manager) is in INACTIVE state

Could someone help me to find where I had went wrong...

I am able to create cluster between queue managers from windows, the problem occurs while adding z/OS queue manager.

I have read that SYSTEM.CLUSTER.COMMAND.QUEUE depth should be zero, but in z/OS queue manager it is not zero.

Queue depth is not 0 in SYSTEM.CLUSTER.COMMAND.QUEUE & SYSTEM.CLUSTER.TRANSMIT.QUEUE.

I do not know whether there is any resource shortage problems regarding z/OS queue manager.

Earlier I was able to add z/OS queue manager as partial repositories with windows queue managers as full repositories. But all those times I had removed z/OS queue managers from those cluster without following proper steps. I had deleted all those windows queue managers also, so those queue managers do not exist now. But still I can see details of that cluster in z/OS when i give dis clusqmgr(*) command.

Is it because my cluster cache is full?

Is this is due to port issue?
Could you please tell how to define a cluster-sender and cluster-receiver channel and the points that we need to keep in mind while creating it.

I had gone through almost most of the posts in this forum but I was unable to find anything exactly related to my problem.
Back to top
View user's profile Send private message
fjb_saper
PostPosted: Thu May 29, 2014 6:01 am    Post subject: Reply with quote

Grand High Poobah

Joined: 18 Nov 2003
Posts: 20696
Location: LI,NY

Did you even read the Cluster manual.
The definition of your cluster channels shows us that you have to go back to the drawing board and start with 0.
And please use the naming conventions! Naming conventions are here to make our life easier...

Don't just go through the posts in the forum
this link
_________________
MQ & Broker admin
Back to top
View user's profile Send private message Send e-mail
Prithviraj
PostPosted: Thu May 29, 2014 9:12 am    Post subject: Reply with quote

Apprentice

Joined: 29 May 2014
Posts: 30

I do not understand what is wrong in my channel definition.

My z/OS ip address is 172.16.100.200 and queue manager port is 1400.
My windows ip address is 172.16.111.222 and queue manager listener port is 1414(I have tried with other port numbers also)


For cluster-receiver definition in z/OS conname should be 172.16.100.200(1400)

For cluster-receiver definition in windows conname should be 172.16.111.222(1414).

Moreover I tried adding z/OS queue manager to cluster using WebSphere MQ Explorer where all channels are created automatically but still z/OS queue manager is not getting added to the cluster.

Please let me know what is wrong with my cluster configuration.
Back to top
View user's profile Send private message
JosephGramig
PostPosted: Thu May 29, 2014 9:45 am    Post subject: Reply with quote

Grand Master

Joined: 09 Feb 2006
Posts: 1230
Location: Gold Coast of Florida, USA

First, all "explicit" cluster sender channels should only point to one or the other Full Repository (cluster receiver channel). That means the two Windows Qmgrs you are using for Full Repositories will only explicitly point to one another.

Let's say your windows Qmgrs are called FR1 and FR2.

Alter these qmgrs to have REPOS(NCLUSTER).

FR1 needs a cluster receiver called TO.FR1 with the conname pointed to the listener on itself.
FR2 needs a cluster receiver called TO.FR2 with the conname pointed to the listener on itself.

FR1 needs a cluster sender called TO.FR2 with the conname pointed to the listener on FR2.
FR2 needs a cluster sender called TO.FR1 with the conname pointed to the listener on FR1.

Now FR1 and FR2 can send back and forth.

Next, inject the z/OS Qmgr (let's call that P1QM).

P1QM needs a cluster receiver called TO.P1QM with the conname pointed to the listener on itself.
P1QM needs a cluster sender called TO.FR1 with the conname pointed to the listener on FR1. This injects P1QM into the cluster and instructs the FRs to build automatic channels back to P1QM.

Note that nowhere did I mention FROM.<QmgrName> for a channel name when talking MQ clustering.

This is outlined in more detail in the Clustering section of the Info Center (which you need to read much more carefully).

Consider naming cluster channels as <ClusterName>.<QmgrName> instead of TO.<QmgrName>. This will ensure unique channel names per Qmgr/Cluster which is easier for the inexperienced folks.
Back to top
View user's profile Send private message AIM Address
Prithviraj
PostPosted: Thu May 29, 2014 10:20 am    Post subject: Reply with quote

Apprentice

Joined: 29 May 2014
Posts: 30

hai JosephGramig,

First of all , let me thank you for your detailed explanation. I appreciate you patience in explaining this much.

Now let us come to topic,

If we have two queue managers, cluster-sender channel definition of one queue manager will be same as cluster receiver channel definition of the other, right?

Similarly, cluster-sender channel definition of partial repos queue manager will be same as cluster-receiver channel definition of one of full repos queue manager.

If this is true, I have defined my channels like that and it is supposed to work.

Moreover I tried it using MQ Explorer where we need not want to create any channels manually as it is created automatically.

In my case, cluster sender channel from partial repos queue manager to full repos queue manager is running, but cluster-receiver channel in partial repos is always in inactive state.

when i give clusqmgr(*) in windows, I am able to see all channels except windows to z/OS channel, ie z/OS cluster- receiver channel or windows cluster sender channel to z/OS.

Note: my partial repos queue manager is in z/OS
Back to top
View user's profile Send private message
Inisah
PostPosted: Thu May 29, 2014 11:16 am    Post subject: Reply with quote

Apprentice

Joined: 21 Mar 2014
Posts: 44

You have not understood the concept clearly Prithviraj. Please try to read the above posts a lil more clearly


Windows - Sender to REPOS and not to Z/OS
Receiver to itself

Z/OS - Sender to REPOS and not to Windows directly
Receiver to itself
Back to top
View user's profile Send private message
Prithviraj
PostPosted: Thu May 29, 2014 9:58 pm    Post subject: Reply with quote

Apprentice

Joined: 29 May 2014
Posts: 30

Hi all,

Thanks for everyone who have replied to my post.

After reading your post, I came to know that there is something wrong in my concept but unfortunately I am unable to figure it out.

Help me in defining my cluster-channels,

My z/OS ip address is 172.16.100.200 and queue manager port is 1400.

My windows m/c ip address is 172.16.111.222 and have two queue managers listening to ports 1421 and 1422.

I need two windows queue mangers as full repos and one z/OS queue manager as partial repos.

please define needed cluster channels required with their connection name..and help me to understand the concept..
Back to top
View user's profile Send private message
exerk
PostPosted: Thu May 29, 2014 11:20 pm    Post subject: Reply with quote

Jedi Council

Joined: 02 Nov 2006
Posts: 6339

Hopefully the following will be a little more illuminating. Note, all names used are purely example only...

Windows Full Repository 1
Code:
ALTER QMGR REPOS('CLUSA')

DEFINE CLUSRCVR(CLUSA.WFR1) +
            CONNAME('WFR1(1421)') +
            CLUSTER('CLUSA') +
            .

DEFINE CLUSSDR(CLUSA.WFR2) +
            CONNAME('WFR2(1422)') +
            CLUSTER('CLUSA') +
            .


Windows Full Repository 2
Code:
ALTER QMGR REPOS('CLUSA')

DEFINE CLUSRCVR(CLUSA.WFR2) +
            CONNAME('WFR2(1422)') +
            CLUSTER('CLUSA') +
            .

DEFINE CLUSSDR(CLUSA.WFR1) +
            CONNAME('WFR1(1421)') +
            CLUSTER('CLUSA') +
            .


z/OS Partial Repository
Code:
DEFINE CLUSRCVR(CLUSA.ZPR1) +
            CONNAME('ZPR1(1400') +
            CLUSTER('CLUSA') +
            .

DEFINE CLUSSDR(CLUSA.WFR1) +
            CONNAME('WFR1(1421') +
            CLUSTER('CLUSA') +
            .


It is better to use DNS names when clustering queue managers as it's a lot less painful should IP Addresses be changed.
_________________
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
View user's profile Send private message
Prithviraj
PostPosted: Thu May 29, 2014 11:44 pm    Post subject: Reply with quote

Apprentice

Joined: 29 May 2014
Posts: 30

Hi everyone,

I have went through all the available MQ manuals and most of the forums about clustering, so please do not tell that I haven't reffered or studied.

You all have given me detailed description about how to define channels but none have told where I went wrong. I have defined my channels as you have said, but still it is not working.

When we use MQ explorer for clustering, all channel definitions will be done automatically.

when we have 3 queue managers with 2 full repos and 1 partial repos,
two queue managers (1 partial and 1 full repos) will have same sender channel definitions as that of other full repos queue manager's cluster-receiver channel, right?

then clssdr of one full repos queue manger will be exact same of clsrcvr of other full repos qmanager, right?

Connection name of cluster-receiver channel of partial repos queue manager will be its ip address(it's listening port).

i have successfully created cluster b/w full repos queue manager in windows, but not able to add z/OS queue manager as partial repos.

Is anything wrong in what I have written above?

If I am correct, what are the other thing I have to check?

Please let me know.

[/img]
Back to top
View user's profile Send private message
exerk
PostPosted: Fri May 30, 2014 4:40 am    Post subject: Reply with quote

Jedi Council

Joined: 02 Nov 2006
Posts: 6339

I am making the assumption here that you have ensured there is connectivity between the servers, e.g. ping, and that there are no channel exits etc. on either side, so I suggest you define a SDR/RCVR channel pair, and associated XMITQ, in one of the Windows queue managers and see if the SDR channel will start or produce errors.
_________________
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
View user's profile Send private message
bruce2359
PostPosted: Fri May 30, 2014 5:09 am    Post subject: Reply with quote

Poobah

Joined: 05 Jan 2008
Posts: 9394
Location: US: west coast, almost. Otherwise, enroute.

Here is a good slide-show on how to set up a cluster: http://www-01.ibm.com/support/docview.wss?uid=swg27009247&aid=1
_________________
I like deadlines. I like to wave as they pass by.
ב''ה
Lex Orandi, Lex Credendi, Lex Vivendi. As we Worship, So we Believe, So we Live.
Back to top
View user's profile Send private message
hughson
PostPosted: Fri May 30, 2014 5:44 am    Post subject: Re: Not able to add z/OS queue manager to cluster Reply with quote

Padawan

Joined: 09 May 2013
Posts: 1914
Location: Bay of Plenty, New Zealand

Prithviraj wrote:
AMQ8441: Display Cluster Queue Manager details.
CLUSQMGR(SYSTEM.TEMPQMGR.172.16.100.200(1400))
CHANNEL(FROM.WINDOWS12) CLUSTER(NCLUSTER)

When you see this SYSTEM.TEMPQMGR you know that the initial clustering flows have not completed successfully. You should check for error messages which say why, especially since you have checked that your channels are successfully running already (which can be the other cause of incomplete cluster initialisation).

Prithviraj wrote:
Queue depth is not 0 in SYSTEM.CLUSTER.COMMAND.QUEUE & SYSTEM.CLUSTER.TRANSMIT.QUEUE.

When you see non-zero curdepths on these queues it tells you that the repository task/process is not operating correctly. You should check for error messages which say why.

So, please can you check the CHINIT address space joblog of your z/OS queue manager for clustering error messages, those will be the CSQX4* ones. Those will help to enlighten us to your plight.

Cheers
Morag
_________________
Morag Hughson @MoragHughson
IBM MQ Technical Education Specialist
Get your IBM MQ training here!
MQGem Software


Last edited by hughson on Fri May 30, 2014 6:12 am; edited 1 time in total
Back to top
View user's profile Send private message Visit poster's website
JosephGramig
PostPosted: Fri May 30, 2014 6:01 am    Post subject: Reply with quote

Grand Master

Joined: 09 Feb 2006
Posts: 1230
Location: Gold Coast of Florida, USA

Prithviraj,

All channels that are to communicate with one another must have matching names (they are pairs).

Please don't use IP address' anywhere. Always use DNS names and if there isn't a DNS name to resolve to, add something to the /etc/hosts file. In the CONNAME in addition to the DNS name, always specify the port number (because some igmo will change the default and mess you up).

If you have been defining cluster objects and they are not working, you must first get rid of all vestiges of those objects before getting anything new to work. This is not always as straight forward as it may seem. MQ clustering will automatically define objects for you (the cluster magic). Those automagic objects will not go away automagically (well, maybe in 90 days). You must know how to use a big ace hammer to kill them. When you think you have removed all cluster definitions, use REFRESH CLUSTER(<ClusterName>) REPOS(YES) to dump all automagic definitions and even then it may take a restart of the Qmgr to truly get it clean. BTW, if you ever think you need to issue REFRESH CLUSTER at a full repository, you have hopelessly messed up.

Where did you go wrong:

  1. You didn't get the Full Repositories defined correctly (as best I can tell from your post)
  2. You used IP addresses in your CONNAMEs
  3. You used poor channel names
  4. You miss understood how channels work
  5. You asked for somebody on the internet to tell you where you went wrong!
  6. When what you do goes wrong, it means you didn't understand what you read in the manual (and random posts on the internet are not a substitution for the manual but a dumbed down version of it that has a significant lost of information)
  7. You didn't read the manual enough (when you can quote it more than the most religious book you know or can think of, you are getting close)


So why didn't anybody correct your definitions? Well, you didn't give enough information to correct the definitions. We couldn't figure out what you were doing, but we could see it was nowhere close to correct or what the manual tells you to do. So we gave you correct procedures and then we even gave you correct commands (well, a good guess but our hands are not on your keyboard so...).

I can be confident you are offended by my comments and hate them.

Finally, read the manual 1000 more times and ask for forgiveness.
Back to top
View user's profile Send private message AIM Address
Prithviraj
PostPosted: Sat May 31, 2014 6:53 am    Post subject: Reply with quote

Apprentice

Joined: 29 May 2014
Posts: 30

Hi JosephGramig,

thanks for your feedback, I will surely go through manuals.

I created new cluster using MQ explorer, with proper naming conventions, still it is not working.

As 2 weeks before I was successful in creating a cluster b/n windows and z/OS, I think it is a problem with z/OS queue manager.

I will restart z/OS MQ and will try again.

Thanks everyone for your valuable suggestions.

Before concluding, do anyone have any other suggestions regarding my problem..?

Any reply will be appreciated.
Back to top
View user's profile Send private message
bruce2359
PostPosted: Sat May 31, 2014 7:02 am    Post subject: Reply with quote

Poobah

Joined: 05 Jan 2008
Posts: 9394
Location: US: west coast, almost. Otherwise, enroute.

Why are you restarting z/OS MQ? It will likely not help.
_________________
I like deadlines. I like to wave as they pass by.
ב''ה
Lex Orandi, Lex Credendi, Lex Vivendi. As we Worship, So we Believe, So we Live.
Back to top
View user's profile Send private message
Display posts from previous:   
Post new topic  Reply to topic Goto page 1, 2  Next Page 1 of 2

MQSeries.net Forum Index » Clustering » Not able to add z/OS queue manager to cluster
Jump to:  



You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
Protected by Anti-Spam ACP
 
 


Theme by Dustin Baccetti
Powered by phpBB © 2001, 2002 phpBB Group

Copyright © MQSeries.net. All rights reserved.