| Author |
Message
|
| jcaron |
 Posted: Wed Apr 05, 2006 12:25 pm Post subject: QFlex & QPasa - AT THE SAME TIME |
 |
|
Apprentice
Joined: 02 Jun 2005
Posts: 30
Location: Indianapolis, IN
|
I was just wondering if anyone has used QPasa & QFlex in the same environment at the same time. We use QPasa for monitoring and alerts and I wanted to use QFlex for gathering statistics like throughput (since it's fairly simple to setup for the kind of data I want). However, I realized that if both of them are using 'reset queue statistics' for gathering data, that this might cause a problem with QPasa's ability to find problems and report about them, as well as throw off the statistics that QFlex gathers. Does anyone have any experience running these two tools together, should I be worried about contention, or is this much ado about nothing?
I'm planning on calling Netflexity about this shortly, if no one knows the answer I mave have one for everyone soon enough  |
|
| Back to top |
|
 |
| jcaron |
| Posted: Wed Apr 05, 2006 12:35 pm Post subject: |
|
|
Apprentice
Joined: 02 Jun 2005
Posts: 30
Location: Indianapolis, IN
|
| Well, according to the folks at Netflexity, QFlex does in fact do 'reset queue statistics' to get its data... time to see if this is what QPasa does as well. Would seem awfully silly since it uses agents for it to collect the data that way, but who knows?? |
|
| Back to top |
|
|
| jefflowrey |
| Posted: Wed Apr 05, 2006 12:39 pm Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
Everyone does reset queue statistics to gather data.
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Wed Apr 05, 2006 12:40 pm Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
Why would you want 2 different tools for the same task -- concurrently?
QPasa gets us the statistics fine. It works out of bandwidth. And yes it will reset the statistics.
Enjoy 
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| jcaron |
| Posted: Wed Apr 05, 2006 12:51 pm Post subject: |
|
|
Apprentice
Joined: 02 Jun 2005
Posts: 30
Location: Indianapolis, IN
|
Yup, confirmed by the folks at mqsoftware... so lame.
Long story about why I wanted to use both, not really worth delving into believe me.
Thanks for the replies. |
|
| Back to top |
|
|
| jcaron |
| Posted: Wed Apr 05, 2006 12:53 pm Post subject: |
|
|
Apprentice
Joined: 02 Jun 2005
Posts: 30
Location: Indianapolis, IN
|
| BTW, statistics can also be gathered via channel exits, so 'reset queue statistics' is not the only way... |
|
| Back to top |
|
|
| jefflowrey |
| Posted: Wed Apr 05, 2006 1:56 pm Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
| jcaron wrote: |
| BTW, statistics can also be gathered via channel exits, so 'reset queue statistics' is not the only way... |
Explain to me how one gathers the dequeue rate of a queue by using a channel exit.
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| jcaron |
| Posted: Wed Apr 05, 2006 5:00 pm Post subject: |
|
|
Apprentice
Joined: 02 Jun 2005
Posts: 30
Location: Indianapolis, IN
|
| jefflowrey wrote: |
| Explain to me how one gathers the dequeue rate of a queue by using a channel exit. |
Not ALL stats, but some - ask the folks at mqsoftware how StatWatch works... especially since it needs to coexist with QPasa. It would be a really bad thing if a company sold two products for MQ that conflicted with each other's operation. |
|
| Back to top |
|
|
| jefflowrey |
| Posted: Thu Apr 06, 2006 4:47 am Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
| jcaron wrote: |
| Not ALL stats, but some - |
Why would "RESET QUEUE STATISTICS" gather any statistics other than those involving queues?
As far as I know, there is no way to get the information that Reset Queue Statistics produces, other than issuing Reset Queue Statistics. Well, I suppose one could incur the overhead of using an API exit, and record the data yourself - but then you might as well gather all the metrics about MQ you want that way. And you've just re-invented the wheel.
Some of the metrics that WebSphere MQ produces can be gathered in a non-destructive manner. Tools that gather those metrics can coexist successfully - as long as they only gather those metrics.
Some metrics, like the ones exposed through Reset Queue Statistics and any of the Event Message based metrics, can only be gathered destructively. Consequently, any tools that gather those metrics will conflict with other tools that gather those metrics.
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| jcaron |
| Posted: Thu Apr 06, 2006 4:55 am Post subject: |
|
|
Apprentice
Joined: 02 Jun 2005
Posts: 30
Location: Indianapolis, IN
|
| jefflowrey wrote: |
| Why would "RESET QUEUE STATISTICS" gather any statistics other than those involving queues? |
It wouldn't and I didn't say it would. I said that a channel exit could get you some stats (like those captured by StatWatch), just not all that RESET can.
| jefflowrey wrote: |
| As far as I know, there is no way to get the information that Reset Queue Statistics produces, other than issuing Reset Queue Statistics. Well, I suppose one could incur the overhead of using an API exit, and record the data yourself - but then you might as well gather all the metrics about MQ you want that way. And you've just re-invented the wheel. |
Agreed, but if someone like mqsoftware has already reinvented the wheel for us, then so be it.
| jefflowrey wrote: |
| Some of the metrics that WebSphere MQ produces can be gathered in a non-destructive manner. Tools that gather those metrics can coexist successfully - as long as they only gather those metrics. Some metrics, like the ones exposed through Reset Queue Statistics and any of the Event Message based metrics, can only be gathered destructively. Consequently, any tools that gather those metrics will conflict with other tools that gather those metrics. |
Couldn't agree more. |
|
| Back to top |
|
|
| jefflowrey |
| Posted: Thu Apr 06, 2006 5:04 am Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
| jcaron wrote: |
| Agreed, but if someone like mqsoftware has already reinvented the wheel for us, then so be it. |
The development costs far outweigh the benefits of allowing your competitor's tools to coexist.
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| jcaron |
| Posted: Thu Apr 06, 2006 5:09 am Post subject: |
|
|
Apprentice
Joined: 02 Jun 2005
Posts: 30
Location: Indianapolis, IN
|
| jefflowrey wrote: |
| The development costs far outweigh the benefits of allowing your competitor's tools to coexist. |
Truth, but they don't outweigh the benefits mqsoftware gets by having their own software being able to coexist (QPasa and StatWatch) - selling two products to a customer is always better than selling just one. |
|
| Back to top |
|
|
| jefflowrey |
| Posted: Thu Apr 06, 2006 5:14 am Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
| jcaron wrote: |
| jefflowrey wrote: |
| The development costs far outweigh the benefits of allowing your competitor's tools to coexist. |
Truth, but they don't outweigh the benefits mqsoftware gets by having their own software being able to coexist (QPasa and StatWatch) - selling two products to a customer is always better than selling just one. |
Only if you're a salesperson. 
Unless they've completely changed since version 3, though, QPasa uses reset queue statistics just like everyone else does.
QNami! on the other hand...
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
|
|
