| Author |
Message
|
| lancelotlinc |
 Posted: Wed Jun 22, 2011 6:51 am Post subject: |
 |
|
Jedi Knight
Joined: 22 Mar 2010
Posts: 4941
Location: Bloomington, IL USA
|
It seems to me the problem you face is caused by the unneeded complexity of the underlying OS processes you have in place. I would solve the problem by rebuilding from scratch to a simple active-active configuration and abandon the active-passive configuration.
Simplicity is a wonderful solution to many problems. Unnecessary complexity drives unecessary cost and unecessary recovery times.
_________________
http://leanpub.com/IIB_Tips_and_Tricks
Save $20: Coupon Code: MQSERIES_READER |
|
| Back to top |
|
 |
| mqjeff |
| Posted: Wed Jun 22, 2011 6:51 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
If it is in a zone, then the errors you are seeing are suggestive of a configuration issue where you have not installed MQ into the right zone.
Has this passive node ever worked? If not, then that is doubly-suggestive of an install issue.
If it has worked at some time in the past, then you need to identify that point in time and determine the oldest point in time after that when it stopped working, and then examine your change management records of all the changes that occurred between those two dates. |
|
| Back to top |
|
|
| Vitor |
| Posted: Wed Jun 22, 2011 7:12 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| lancelotlinc wrote: |
| I would solve the problem by rebuilding from scratch to a simple active-active configuration and abandon the active-passive configuration. |
Presupposing whatever went wrong before doesn't go wrong in the new build.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| Vitor |
| Posted: Wed Jun 22, 2011 7:14 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| mqjeff wrote: |
| Has this passive node ever worked? If not, then that is doubly-suggestive of an install issue. |
Perhaps more specifically has anything changed in the zone configuration? Running WMQ in a Solaris zone is fiddley at best. What kind of zone install are you using?
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| Vitor |
| Posted: Wed Jun 22, 2011 7:16 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| lancelotlinc wrote: |
| Simplicity is a wonderful solution to many problems. Unnecessary complexity drives unecessary cost and unecessary recovery times. |
This also assumes that the set up in question isn't protected by an active / passive set up because it's production. Clearly it's an environment that's deemed worthy of such expensive protection, and if migrating to WMBv7 is described by the OP as a "struggle", tearing down an important box and rebuilding it may not be as easy as you make it sound.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| lancelotlinc |
| Posted: Wed Jun 22, 2011 7:24 am Post subject: |
|
|
Jedi Knight
Joined: 22 Mar 2010
Posts: 4941
Location: Bloomington, IL USA
|
I agree that production systems need protection. I advocate that the expense incurred by active-passive configurations are wasted money. We've spent all this money and still cannot protect the production system. Time-to-recover of weeks or months is not acceptable.
_________________
http://leanpub.com/IIB_Tips_and_Tricks
Save $20: Coupon Code: MQSERIES_READER |
|
| Back to top |
|
|
| Vitor |
| Posted: Wed Jun 22, 2011 7:34 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| lancelotlinc wrote: |
| I advocate that the expense incurred by active-passive configurations are wasted money. |
We all know you advocate that. It's not an answer to the OP's question.
| lancelotlinc wrote: |
| Time-to-recover of weeks or months is not acceptable. |
No it isn't. Which is why you test these things, as the OP may have done at the end of May & discovered there's a problem. You are again assuming that the OP is actually trying to fail over a live system (and if production's been down since the end of May the site manager has the patience of Job...)
Or the work at the end of May could have been an install of any of the products or (as I postulate above) a change to the zone configuration. This ties back to the excellent advice of checking changes between the last time it worked and now, or if it's never working in it's current configuration checking installation & said configuration.
Especially if the change was to the zone(s).
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| lancelotlinc |
| Posted: Wed Jun 22, 2011 7:40 am Post subject: |
|
|
Jedi Knight
Joined: 22 Mar 2010
Posts: 4941
Location: Bloomington, IL USA
|
| Vitor wrote: |
| lancelotlinc wrote: |
| I advocate that the expense incurred by active-passive configurations are wasted money. |
We all know you advocate that. |
Oh. I  thought I was not being outspoken enough.
_________________
http://leanpub.com/IIB_Tips_and_Tricks
Save $20: Coupon Code: MQSERIES_READER |
|
| Back to top |
|
|
| mqjeff |
| Posted: Wed Jun 22, 2011 7:49 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
It really doesn't matter how outspoken you choose to be.
The math on active-active absolutely does not automatically add up to either a cheaper or a better solution. This is factual, and demonstratable, and I will not repeat the full explanation for it here. But as a simple point, there is a fixed capacity that is required in a given situation. You are required to purchase enough licenses and hardware to meet this capacity, and ensure this capacity is available at all times under the SLA. It is then merely a matter of identifying tradeoffs in terms of manageability, recoverability and scalability that meet business requirements to decide how many buckets to put the capacity into. |
|
| Back to top |
|
|
| Vitor |
| Posted: Wed Jun 22, 2011 7:54 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| lancelotlinc wrote: |
| Oh. I thought I was not being outspoken enough. |
Don't worry - you're being plenty loud and clear.
Your strategy has not been universally adopted for other reasons.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| kris.pilaet |
| Posted: Wed Jun 22, 2011 9:14 am Post subject: |
|
|
Newbie
Joined: 21 Jun 2011
Posts: 8
|
Hello all,
first of all, I'm glad my post is being discussed 
But I think I need to clarify a bit our situation.
Concerning myself, I'm an MQ & Broker administrator. To be honest, just since 1/1 of this year, but my knowledge of those is okay (and growing - learning).
The OS, that's the responsibility of another team. We're a (too) large company. Don't misread me, I'm not taking my umbrella, and not saying "that's not my fault / problem".
My understanding of the OS, zones, clusters, ... is not yet good enough (working on it).
I'm just trying to find a solution...
That said, I'm gonna upset some people now 
I'm facing this problem in Test.
Node A has worked long and good for, hell, long.
Node B is equal to A, only, now our only Test-environment.
Acceptance & Production has the same setup, and working good.
If suggestion is to look into changes on zone-level, I accept that. Though, people say there equal on both nodes. I'll pick up the issue.
Is this the only suggestion (so far) you can give me? Or are there others?
-> Still I do agree that this should be it, since MQ & Broker versions / install / setup issues are indeed the same. |
|
| Back to top |
|
|
| Vitor |
| Posted: Wed Jun 22, 2011 9:21 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| kris.pilaet wrote: |
| first of all, I'm glad my post is being discussed |
I'm sorry you kind of got hijacked there. I was more facilitator than moderator there.... 
But I think I need to clarify a bit our situation.
| kris.pilaet wrote: |
| Acceptance & Production has the same setup, and working good. |
I'm sure they have the same designed architecture, and I'm certain you've been assured that the boxes are configured identically. I view this assurance the same way I view a network team telling me that the channels suddenly going into retry isn't a network issue.
| kris.pilaet wrote: |
| If suggestion is to look into changes on zone-level, I accept that. |
It's an obvious place to start looking. I'm also unclear on what's brought the issue to light. Did you try (unsuccessful) to failover in May? Were you testing the failover in May? Were you settting up the failover in May?
I think these important questions may have got lost in the earlier discussion.
To summarize:
Have you ever successfully failed over?
If so, when?
Why did you start looking into this problem in May?
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| kris.pilaet |
| Posted: Wed Jun 22, 2011 9:34 am Post subject: |
|
|
Newbie
Joined: 21 Jun 2011
Posts: 8
|
@Vitor, no apologies needed, it was instructive.
We've failed over (unplanned) with success a zillion times (I don't know how many times) in the past
The problem has been brought to light when stopping the broker for the installation of another product (tivoli Omegamon) and starting it again (where the starting brought the described problems).
So we've switched to B, all ok (product was installed there as well).
When I have the chance, I switch now from B to A to test stuff out and get it working again. But at the moment, that hasn't brought me much joy...
Hope this clarisfies some things? |
|
| Back to top |
|
|
| Vitor |
| Posted: Wed Jun 22, 2011 9:44 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| kris.pilaet wrote: |
| Hope this clarisfies some things? |
Very much so.
So you've got a working active / passive that's happy as the day is long, goes from one node to another with the grace of a dancer, you install Omegamon on both nodes and encounter the described problem but only on 1 node.
This points to the install / configuration on the failing node. Especially if Omegamon is running correctly on node B. And there's been no change to the WMQ/WMB software on either node.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| kris.pilaet |
| Posted: Wed Jun 22, 2011 10:09 am Post subject: |
|
|
Newbie
Joined: 21 Jun 2011
Posts: 8
|
@Vitor I'll pick it up the coming days at work and discuss the things said here.
I'll keep this post updated for the outcome of whatever that's gonna come.
Already thanks for the input |
|
| Back to top |
|
|
|
|
