all numbers are based on client location - so plan to run on client computer
motty
---In vantage@yahoogroups.com, <scott.litzau@...> wrote:Digging up an old thread.
Just curious, are these number being reported run at the server or the client.
I am just digging into all the performance tuning and see bad network numbers at the client but great numbers on my app server.
Scott
From: vantage@yahoogroups.com [mailto:vantage@yahoogroups.com] On Behalf Of too_much_hg
Sent: Thursday, March 21, 2013 9:44 AM
To: vantage@yahoogroups.com
Subject: [Vantage] Re: E9 Performance Diagnostic Tool
Thanks to all for the feedback.
I agree, my numbers are pretty good but I don't understand why an E9 configuration would have double "Server" numbers.
I am still testing and plan to run some IOMETER tests.
I have restarted the services and rebooted the server but always get the same results.
I also agree that splitting SQL from your APP Server is slower than combining them. We have 80 Full Vantage Users and 40 MES Users we run it on a combined configuration.
Regarding the Clients hanging up, that is almost always network latency. I also am unaware of any configuration that would allow for longer timeouts.
If you are seeing this often I recommend you look carefully at all your switch port configurations. As I have posted in other threads it is hugely important that you don't "mismatch" the settings on the switch port and the connected machine (Server or Client).
I always Force my Server NIC settings as well as my Switch Port so that I know they match and aren't repeatedly negotiating in Auto.
Clients I just let the Auto negotiate.
If you have your ports "Teamed-Trunked" make sure that you are choosing a protocol that is correct for your switch, you will find this in the advanced section of the NIC TEAM configuration.
I have to set my switch teams to use (Routing Based on IP-Hash)for VMware, and (Switch Assisted Load Balancing) in our HP TEAMING.
This is a bit unusual and certainly not the default when creating trunks which usually will choose FIP or 8.02.3ad or default or some other scheme.
Read up on your switch and test the different options, they don't usually fail, they will just not work quite right. I would rather things failed because when they "almost" work correctly it is easy to think you have it right when you really don't !
Neil
--- In vantage@yahoogroups.com, "mikethorne1980" <mikethorne1980@...> wrote:
>
>
>
> This is the syntax for restarting the appservers (905.xxx), granted you will need to modify it to work for your environment:
> ---------------------------------------------------
> @echo off
> echo Restarting Appservers
> set DLC=D:\PROGRESS\oe102a
> set PATH=%DLC%\BIN;%PATH%
> set LIB=%DLC%\LIB;%LIB%
>
> echo STOPPING LOCAL APPSERVERS AND DATABASE
> call asbman -name Epicor905ProcessServer -stop
> call asbman -name Epicor905TaskAgent -stop
> call asbman -name Epicor905 -stop
> REM call dbman -database Epicor905 -stop
>
> ping 127.0.0.1 -n 30 > nul
>
> echo STARTING LOCAL APPSERVERS AND DATABASE
> REM call dbman -database Epicor905 -start
> call asbman -name Epicor905 -start
> call asbman -name Epicor905TaskAgent -start
> call asbman -name Epicor905ProcessServer -start
> -----------------------------------------------------
>
> It is possible to stop/start the database, but we decided to not use that part of the script.
>
> As far as the timeout setting, I am unaware of that setting but would be interested in implementing it.
>
> -Mike
>
>
> --- In vantage@yahoogroups.com, "jseeman21" <jseeman@> wrote:
> >
> > Hello,
> >
> > Is it possible to obtain a copy of your batch job as we would like to implement the same. In addition I heard there is a Time Out setting that can be set to prevent Epicor from crashing instantly when there is a network blip? In fact we find that epicor crashes instantly with any network interruption.
> >
> > Look forward to hearing from you.
> >
> > Cheers..
> > J Seeman
> >
> > --- In vantage@yahoogroups.com, "mikethorne1980" <mikethorne1980@> wrote:
> > >
> > >
> > >
> > > Neil,
> > >
> > > It was stated in one of the Epicor best practices documents, but restarting the three appservers via a batch job every night significantly improved our network time (not too sure how). We went from around .08-1.2 to .45 on Win7 clients and .65 on WinXP. Is that something you have implemented?
> > >
> > > By the way, the numbers you are seeing are very good and not something I would think you need to worry about. The tool does state that on a LAN network time should be below .4 which you are. I hope to be able to get down to where you are eventually. Curious, have you run the tool on your clients at all?
> > >
> > > Thanks,
> > > -Mike
> > >
> > >
> > > --- In vantage@yahoogroups.com, "too_much_hg" <neil_willet@> wrote:
> > > >
> > > > I am running in VMware with a Fiber Channel SAN back end.
> > > >
> > > > HP DL380 G8 Host Servers, plenty of horsepower and RAM.
> > > >
> > > > I have a virtual copy of my current physical V8 Server and that actually outperforms the physical server (slightly).
> > > >
> > > > I have this E9 VM running on the same Host and get the results listed below. That is why I am wondering about configuration tweaks vs expected results.
> > > >
> > > > Is E9 just slower ?
> > > >
> > > > This is not apples to apples because my V8 Server OS is 2003 X64 and SQL 2005 .... my E9 Server OS is 2008 R2 X64 and SQL 2008.
> > > >
> > > > I would really like to get the same numbers as my virtual V8 box.
> > > >
> > > > In the diagnostic tool the Network numbers are similar but the server number is roughly double. Trying to put my finger on that part.
> > > >
> > > > I am open to suggestions for things to check or tweak.
> > > >
> > > > Thanks for the help,
> > > > Neil
> > > >
> > > > --- In vantage@yahoogroups.com, Joshua Giese <jgiese@> wrote:
> > > > >
> > > > > Beyond the best practices documents I don't really know what else to tune. Never have experimented myself to find out what effect certain changes might have on the system. I've learned to accept that with Epicor performance there is no "Smoking Gun" but more of a blood trail... Closest thing I've gotten to a smoking gun is going from HDDs to SSDs in addition to some fixes to bad customizations. What kind of equipment do you have?
> > > > >
> > > > > reply Signature
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > Joshua Giese
> > > > >
> > > > > CTO
> > > > >
> > > > > 920.437.6400 Ext. 337
> > > > >
> > > > > Site ID: 27450-E905700B2-SQL64
> > > > >
> > > > > ----- Original Message -----
> > > > >
> > > > > From: "too_much_hg" <neil_willet@>
> > > > > To: vantage@yahoogroups.com
> > > > > Sent: Thursday, March 14, 2013 12:03:22 PM
> > > > > Subject: [Vantage] Re: E9 Performance Diagnostic Tool
> > > > >
> > > > >
> > > > >
> > > > > Joshua,
> > > > >
> > > > > Thanks for that info. I have read their documents and have tried a few things but I was hoping for some "real world" advice that might not be contained in their documentation.
> > > > >
> > > > > My system is SQL and it is on the same machine as Epicor. (not split)
> > > > >
> > > > > Any advice like .pf settings or other magic would be appreciated.
> > > > >
> > > > > Thanks,
> > > > > Neil
> > > > >
> > > > > --- In vantage@yahoogroups.com , Joshua Giese <jgiese@> wrote:
> > > > > >
> > > > > > Epicor has some best practices documents that you get get from tech support that list specific adjustments to make on the app servers and if your on SQL within SQL setup and maintenance. About 4-5 documents
> > > > > >
> > > > > >
> > > > > > Joshua Giese
> > > > > >
> > > > > > CTO
> > > > > >
> > > > > > 920.437.6400 Ext. 337
> > > > > >
> > > > > > Site ID: 27450-E905700B2-SQL64
> > > > > >
> > > > > > ----- Original Message -----
> > > > > >
> > > > > > From: "too_much_hg" <neil_willet@>
> > > > > > To: vantage@yahoogroups.com
> > > > > > Sent: Thursday, March 14, 2013 8:47:22 AM
> > > > > > Subject: [Vantage] Re: E9 Performance Diagnostic Tool
> > > > > >
> > > > > >
> > > > > >
> > > > > > We are gettin OK numbers, but I would like to get better on my Epicor 9.04.701 Server when compared to Vantage 8.03.409C
> > > > > >
> > > > > > Server Network
> > > > > >
> > > > > > Physical V8 Server : .12 .28
> > > > > > Virtual V8 Server : .11 .21
> > > > > > Virtual E9 Server : .26 .31
> > > > > >
> > > > > > Why is this double ? ^^^
> > > > > >
> > > > > > Can someone recommend some tuning parameters for the E9 Server ?
> > > > > >
> > > > > > We have about 30 Heavy users, 40 light users and 30 MES users.
> > > > > >
> > > > > > Thanks,
> > > > > > Neil
> > > > > >
> > > > > > --- In vantage@yahoogroups.com , Joshua Giese <jgiese@> wrote:
> > > > > > >
> > > > > > >
> > > > > > > We have about 15 concurrent users at a time on average. To be honest I have not seen a lot of client side speed up from these changes except for in these areas:
> > > > > > > - Crystal Reports generate and transfer faster
> > > > > > > - BO methods are faster by 10ths of seconds only (Could be affected by other hardware changes made more RAM we added, better processors etc)
> > > > > > > - BAQs and Searches run faster by about 10%
> > > > > > > - Screens cache and load faster (Could be affected by other hardware changes made more RAM we added, better processors etc)
> > > > > > >
> > > > > > > One of the only other things from a client perspective that seemed to make an impact was adding an SSD to client computers. Nothing else RAM, CPU, etc seemed to make a significant difference and the SSD only helped screen loads and refreshes. BO method timing stayed about the same and improved again by 10ths of seconds.
> > > > > > >
> > > > > > >
> > > > > > > Joshua Giese
> > > > > > >
> > > > > > > CTO
> > > > > > >
> > > > > > > 920.437.6400 Ext. 337
> > > > > > >
> > > > > > > Site ID: 27450-E905700B2-SQL64
> > > > > > >
> > > > > > > ----- Original Message -----
> > > > > > >
> > > > > > > From: "mikethorne1980" <mikethorne1980@>
> > > > > > > To: vantage@yahoogroups.com
> > > > > > > Sent: Wednesday, March 6, 2013 4:13:29 PM
> > > > > > > Subject: [Vantage] Re: E9 Performance Diagnostic Tool
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > Did you notice a significant client performance increase Josh? Also how many users do you currently have?
> > > > > > >
> > > > > > > --- In vantage@yahoogroups.com , Joshua Giese wrote:
> > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > > Best thing I ever did...
> > > > > > > >
> > > > > > > > Took my Epicor App OS out of a striped RAID array
> > > > > > > >
> > > > > > > > Put SQL and Epicor on the same box with SSD drives.
> > > > > > > >
> > > > > > > > ÃÆ'‚
> > > > > > > >
> > > > > > > > All my times dropped to the posted averages. I found that when SQL and Epicor were on seperate physical boxes that Epicor was always waiting for SQL to get data back to it, but SQL was waiting for input from Epicor because I had aÃÆ'‚ hard drive bottleneck.ÃÆ'‚ It was a vicious circle.ÃÆ'‚ I put a crossover between the two machines, put those two NICs in their own subnet and took the switch out of the equation.ÃÆ'‚ It helped marginally.
> > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > > Joshua Giese
> > > > > > > >
> > > > > > > > CTO
> > > > > > > >
> > > > > > > > 920.437.6400 Ext. 337
> > > > > > > >
> > > > > > > > Site ID: 27450-E905700B2-SQL64
> > > > > > > >
> > > > > > > > ----- Original Message -----
> > > > > > > >
> > > > > > > > From: "timlester67"
> > > > > > > > To: vantage@yahoogroups.com
> > > > > > > > Sent: Friday, March 1, 2013 9:56:20 AM
> > > > > > > > Subject: [Vantage] Re: E9 Performance Diagnostic Tool
> > > > > > > >
> > > > > > > > ÃÆ'‚
> > > > > > > >
> > > > > > > >
> > > > > > > > Curious how much tuning everyone did to obtain the numbers posted. Our server times are much higher than everyone else. We use 1 App server and 1 SQL server, physical boxes. We are seeing the following:
> > > > > > > >
> > > > > > > > Production:
> > > > > > > > Server .60
> > > > > > > > Network .38
> > > > > > > >
> > > > > > > > Test:
> > > > > > > > Server .58
> > > > > > > > Network .37
> > > > > > > >
> > > > > > > > Thanks.
> > > > > > > >
> > > > > > > > Tim Lester
> > > > > > > >
> > > > > > > > --- In vantage@yahoogroups.com , "mseal" wrote:
> > > > > > > > >
> > > > > > > > > before correction server time was just about as after (.15 - .20) but 'network time' was around .55 ++
> > > > > > > > > motty
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > --- In vantage@yahoogroups.com , "jgiese1988" wrote:
> > > > > > > > > >
> > > > > > > > > > Do you recall your numbers before hand and after the upgrade before correcting the issue? What did you see on your new boxes after correcting the issue.
> > > > > > > > > >
> > > > > > > > > > --- In vantage@yahoogroups.com , "mseal" wrote:
> > > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > > > A footnote to `network time' values and VW configuration.
> > > > > > > > > > > About 9 months ago I upgraded all my production host servers on VM to top of the line Dell servers and left the `old' server as part of the DRS configuration so I can move the Epicor server over.
> > > > > > > > > > > Once on the new environment `Server time' values were same as old but network time was x2 (worst).
> > > > > > > > > > > It took a lot of time to figure issue out and main issue was BIOS power profile configuration
> > > > > > > > > > > Bottom line, not sure what Epicor considers `Network time' as it seems to be all calculated on the actual box
> > > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > > > Motty
> > > > > > > > > > >
> > > > > > > > > > > --- In vantage@yahoogroups.com , "jgiese1988" wrote:
> > > > > > > > > > > >
> > > > > > > > > > > > We get something like
> > > > > > > > > > > > .13 server
> > > > > > > > > > > > .48 network
> > > > > > > > > > > >
> > > > > > > > > > > > I wouldn't put a lot of stock in the server time as we have really long disk queues that are killing us on the server.
> > > > > > > > > > > >
> > > > > > > > > > > > --- In vantage@yahoogroups.com , "mseal" wrote:
> > > > > > > > > > > > >
> > > > > > > > > > > > > Thanks John,
> > > > > > > > > > > > > I'm getting
> > > > > > > > > > > > > Server - .15
> > > > > > > > > > > > > Network - .33
> > > > > > > > > > > > >
> > > > > > > > > > > > > I wonder how you consider your overall performance on a 1-5
> > > > > > > > > > > > >
> > > > > > > > > > > > > Motty
> > > > > > > > > > > > >
> > > > > > > > > > > > >
> > > > > > > > > > > > > --- In vantage@yahoogroups.com , John Driggers wrote:
> > > > > > > > > > > > > >
> > > > > > > > > > > > > > There are a lot of variables but it's probably as good a place as any to
> > > > > > > > > > > > > > start.
> > > > > > > > > > > > > >
> > > > > > > > > > > > > > I'm getting:
> > > > > > > > > > > > > > Server: .14
> > > > > > > > > > > > > > Network: .36
> > > > > > > > > > > > > >
> > > > > > > > > > > > > > *John Driggers*
> > > > > > > > > > > > > > **
> > > > > > > > > > > > > > *Chief Data Wrangler*
> > > > > > > > > > > > > > *
> > > > > > > > > > > > > > *
> > > > > > > > > > > > > > *I have an Epicor blog http://usdoingstuff.com/ >. How useful is that?*
> > > > > > > > > > > > > > *
> > > > > > > > > > > > > > *:: 904.404.9233
> > > > > > > > > > > > > > :: waffqle@
> > > > > > > > > > > > > > :: http://www.usdoingstuff.com http://www.usdoingstuff.com/ >
> > > > > > > > > > > > > >
> > > > > > > > > > > > > > *
> > > > > > > > > > > > > >
> > > > > > > > > > > > > > *
> > > > > > > > > > > > > >
> > > > > > > > > > > > > >
> > > > > > > > > > > > > >
> > > > > > > > > > > > > > On Mon, Dec 3, 2012 at 11:36 AM, mseal wrote:
> > > > > > > > > > > > > >
> > > > > > > > > > > > > > > **
> > > > > > > > > > > > > > >
> > > > > > > > > > > > > > >
> > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > For all of us on Epicor 9, wonder if one company's stats of the `Network
> > > > > > > > > > > > > > > Diagnostics' represents a good benchmark for server/dB and client
> > > > > > > > > > > > > > > configuration and performance.
> > > > > > > > > > > > > > > I was hoping Epicor would publish benchmark ranges and I could assess how
> > > > > > > > > > > > > > > good/bad my settings are.
> > > > > > > > > > > > > > > Would that work?
> > > > > > > > > > > > > > > Thanks
> > > > > > > > > > > > > > > Motty
> > > > > > > > > > > > > > >
> > > > > > > > > > > > > > >
> > > > > > > > > > > > > > >
> > > > > > > > > > > > > >
> > > > > > > > > > > > > >
> > > > > > > > > > > > > > [Non-text portions of this message have been removed]
> > > > > > > > > > > > > >
> > > > > > > > > > > > >
> > > > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > >
> > > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > > [Non-text portions of this message have been removed]
> > > > > > > >
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > [Non-text portions of this message have been removed]
> > > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > > [Non-text portions of this message have been removed]
> > > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > [Non-text portions of this message have been removed]
> > > > >
> > > >
> > >
> >
>