|
Home > Archive > BizTalk Server Orchestration > January 2006 > BizTalk 2004 :: too many persistence points, dont know why
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
BizTalk 2004 :: too many persistence points, dont know why
|
|
|
|
Hello,
I posted the other day about persistence becuase I have been getting too
many persistence points in my orchestration when I expected many fewer.
For reference here is the news item:
news://news.microsoft.com/7127ff7bc...s.microsoft.com
I adjusted my orchestration by changing a long running scope to an atomic
transaction. This helped, but I cant understand why I have 3 times more
persistence points than I was expecting.
Why?
I should have 6: 2 atomic scopes, end of orchestration, a long running transaction,
and 2 send shapes. When I open perfmon, select Xlang, chose persistence
points, and run and orchestration instance do do I get 15?? Almost 3 times
more!
The orchestration completes in under 2 seconds, there is no dehydration,
no break points or start orchestrations.
It does have some mappings, calls to the rules engine, sets some xpath variables,
and makes a few decisions. None of which consititute persistence points.
Can someone offer me some ideas as to why I am seeing so many persistence
points? Could it be the perfmon tool interpreting things differently than
the official biztalk documentation would?
MS documentation: http://msdn.microsoft.com/library/d...g_orch_rrdy.asp
Thanks!
BA
http://biztalkia.blogspot.com/
| |
| Stephen W. Thomas 2006-01-13, 9:54 pm |
| I take it you are using the Performance Counters to check the number of
persistence points?
If so, I found that is not accurate and should be used more as a relative
idea of the number of persistence points. i.e if it says 15 and now says 5
you are doing well.
All in all, I would not be too concerned with reducing “every possible
persistence point” unless you are in a high demand environment or already
having performance issues. Make sure you understand the impact of removing
your persistence points in the event of a crash and the Orchestration
restarts.
Hope this helps.
Stephen W. Thomas
http://www.biztalkgurus.com
http://geekswithblogs.net/sthomas/
"BA" wrote:
>
> Hello,
>
> I posted the other day about persistence becuase I have been getting too
> many persistence points in my orchestration when I expected many fewer.
> For reference here is the news item:
>
> news://news.microsoft.com/7127ff7bc...s.microsoft.com
>
> I adjusted my orchestration by changing a long running scope to an atomic
> transaction. This helped, but I cant understand why I have 3 times more
> persistence points than I was expecting.
>
> Why?
>
> I should have 6: 2 atomic scopes, end of orchestration, a long running transaction,
> and 2 send shapes. When I open perfmon, select Xlang, chose persistence
> points, and run and orchestration instance do do I get 15?? Almost 3 times
> more!
>
> The orchestration completes in under 2 seconds, there is no dehydration,
> no break points or start orchestrations.
>
> It does have some mappings, calls to the rules engine, sets some xpath variables,
> and makes a few decisions. None of which consititute persistence points.
>
> Can someone offer me some ideas as to why I am seeing so many persistence
> points? Could it be the perfmon tool interpreting things differently than
> the official biztalk documentation would?
>
> MS documentation: http://msdn.microsoft.com/library/d...g_orch_rrdy.asp
>
> Thanks!
>
> BA
> http://biztalkia.blogspot.com/
>
>
>
| |
|
| Hello Stephen,
That is correct, I am using performance counters, and I am glad to hear its
not accurate (although in general, I wish it was)!
Actually I managed to reduce the number of persistence points from 21 down
to 15, what I expected to see was 6. I did manage to shed a couple.
The performance counters must be counting more writes to the database than
just persistence points. I wish there was something that measured it more
accurately.
Thanks,
BA
http://biztalkia.blogspot.com/
[vbcol=seagreen]
> I take it you are using the Performance Counters to check the number
> of persistence points?
>
> If so, I found that is not accurate and should be used more as a
> relative idea of the number of persistence points. i.e if it says 15
> and now says 5 you are doing well.
>
> All in all, I would not be too concerned with reducing "every possible
> persistence point" unless you are in a high demand environment or
> already having performance issues. Make sure you understand the
> impact of removing your persistence points in the event of a crash and
> the Orchestration restarts.
>
> Hope this helps.
>
> Stephen W. Thomas
> http://www.biztalkgurus.com
> http://geekswithblogs.net/sthomas/
> "BA" wrote:
>
| |
| Marian Drumea [MVP] 2006-01-13, 9:54 pm |
| Hey BA,
You know Heisenberg's uncertainty principle ... On the same idea, if
you have tools to tell everything with 100% accuracy, real-time, then
you end up with a less performant environment and your very results
will possibly be affected by the monitoring. 
Thanks,
Marian Drumea [MVP]
http://www.MarianDrumea.com
|
|
|
|
|