Tuesday, 28 February 2017

Tip from the trenches, interpreting R9698711

Here is a nice gotcha!

Quite often I use the R9698711 to tell me what tables are out of sync between the database and specs, good report.

Here is a problem though, if there is a spec difference – the row count will be 0 – no matter if there are lots of rows.

For example,

Object Name Product Code Record  Count Data Class DS Type Primary Key  (YN) Unique Indexes (In Table) Total Indexes (in Table) Total Indexes (In Specs) Message (Difference between Table and Specification)
F1217Z2 12 0 B I Column(s)  LATT, LONG  in Specs, not found in table
F40051 40 0 B I Column(s)  DMCS, DMCT, OLOCN, OLOTN  in Specs, not found in table
F42I010 42I 0 B I Column(s)  URAB, URAT, URCD, URDT  in Specs, not found in table
F42I015 42I 0 B I Column(s)  BCRC, CRCD, LOTN, OIPR  in Specs, not found in table

 

There are 100000’s of rows in these tables from 9.1 to 9.2.  Interestingly there is no TC provided from JDE to convert the data in them (especially F470371 & F40051).

Of course, if you just do what I do and drop all these tables and create them because they have 0 rows, you’d be making a mistake!

Monday, 27 February 2017

9.2 upgrade running TC’s and getting error on F98950 conversion

 

The first conversion is always the hardest to get going.

I’ve done this 100 times before, perhaps add another 0

I’m running all TC’s locally, because there are hardly any proper ones in the 9.2 upgrade (from 9.1).  I’m also avoiding the production AS/400.

I’m getting the old System – 910 password entry dialog, and I have the following in the logs: (jde.log)

ODB0000164 - DBC:00 [28000][8002][1]

Okay, so on a hunch I start looking around and have seen https://support.oracle.com/epmos/faces/DocContentDisplay?_afrLoop=193155969433737&id=653867.1&_afrWindowMode=0&_adf.ctrl-state=171cvx1f3w_201 which looks really similar, they suggest whacking your current user (in my case JDEUPG) into the [DSPWD] section of the JDE.INI

So I do this (I’m AS/400 and not using security server), and all of my problems go away!

Nice little fix to what could have been a big problem.

Monday, 20 February 2017

ERP analytics webinar

If you are keen on learning more about your JD Edwards installation from a user and performance perspective, this is a must attend session.  

http://myriad-it.com/news/webinarerp-analytics-see-things-differently/

I’ll be covering off the following topics and more:

  1. Evaluate & improve productivity
  2. Create a performance baseline for measuring improvements
  3. Organise your upgrade projects and save time & money while you do it

Date: Wednesday, 1 March 2017
Time: 9am – 10am (AKST) |  10am – 11am (PST)  |  11am – 12pm (MST)  |  12pm – 1pm (CST)  |  1pm – 2pm (EST)

Thursday, 16 February 2017

Friday, 10 February 2017

keep up to date with JD Edwards testing

Have you ever wanted to compare what people do in production (say PD910) and want to make sure that your testing and training environments have run similar applications?

This is really easy if you have out (www.myriad-it.com) plugin google analytics into JD Edwards.

We can create an automated report that emails you all of the apps that have been run in the last day or week in your test environments and email that to you as a CSV.  We can also send you a report of the top 100 apps (by times loaded, or time on page) for your production environment.

A quick vlookup and you are going to know how your test cases are going and how much has been run and tested in your test environment.  This will give you a lot of confidence that your UAT has hit all of the important applications.

Note that this testing can also compare the page load times (as well as the fact that they were loaded), so you can also do some basic performance evaluation too.

Sample CSV:

Page Title application Avg. Document Content Loaded Time (sec) Avg. Document Interactive Time (sec) Avg. Page Download Time (sec) Avg. Page Load Time (sec) Avg. Server Connection Time (sec) Avg. Server Response Time (sec) Pageviews
Work With Addresses P01012_W01012B_ZJDE0001 2.54 2.54 0.16 2.71 0.05 0.41 116
Work With Interactive Versions P983051_W983051I 1.09 1.09 0.05 1.19 0 0.25 85
Paid Time Off P053020_W053020A_ZJDE0006 1.87 1.87 0.2 2.05 0 0.09 63
Employee Profile and Job Information P060116_W060116C_ZJDE0001 1.38 1.38 0.19 1.39 0 0.04 51
Work With Batches P0011_W0011A_ZJDE0001 2.25 2.25 0.26 2.29 0 0.35 43
Safety Scoreboard P54HSPT_S54HSPTA 1.53 1.53 0.1 1.96 <0.01 0.44 38
Select User Define Code P0005S_W0005SB 0.78 0.78 0.1 0.87 0 0.15 38
(not set) P060116_W060116C_ZJDE0001 2.13 2.13 0.13 2.47 0 <0.01 35
Approval Level Revisions - Work With Approval Level Revisions P43008_W43008A_ZJDE0001 0.4 0.4 0.06 0.41 0 0.11 26
Overtime - Rule Set P186401_W186401H_ZJDE0001 1 1 0.2 1.2 0 0.22 24
Work With Self-Service Process Flow Scripts P05400_W05400A 1.6 1.6 0.11 1.65 0 0.26 23
Account Statistical Summary P03B16_W03B16E 2.22 2.22 0.21 2.22 0 0.42 20
Pay Cycle Workbench - Work With Pay Cycle Workbench P07210_W07210A_ZJDE0001 1.37 1.37 0.13 1.41 0 0.75 19
Process Flow Detail P05400_W05400D 1.39 1.39 0.08 1.42 0 0.09 19
Supplier Ledger Inquiry P0411_W0411G_ZJDE0001 3.9 3.9 0.43 3.95 0 0.96 19
Work With Item Master Browse P4101_W4101E_ZJDE0001 2.14 2.14 0.04 2.43 0 0.9 18
Work With Order Activity Rules P40204_W40204B 0.67 0.67 0.09 0.7 0 0.25 18
Work With Orders Awaiting Approval P43081_W43081A_ZJDE0002 0.33 0.33 0.04 0.35 0 0.04 18
My%20System%20Options - User Default Revisions P0085_W0085N 2.36 2.35 0.07 2.51 0 0.59 16
Work With Order Headers P4310_W4310I_ZJDE0001 2.14 2.14 0.09 2.41 0 0.4 16
Order Header P4310_W4310B_TEST 0.71 0.71 <0.01 0.73 0 0.49 14
User Defined Codes - Work With User Defined Codes P0004A_W0004AA_ZJDE0001 2.37 2.37 0.13 2.51 0 0.61 14
Address Book Master Search P0101S_W0101SA 0.46 0.46 <0.01 0.49 0 0.04 13
Approval Level Revisions - Approval Level Revisions P43008_W43008C_ZJDE0001 0.29 0.29 0.06 0.3 0 0.05 13
Interactive Application Search and Select P9860S2_W9860S2A 0.32 0.32 <0.01 0.34 0 0.1 13
    2.06 2.06 0.18 2.18 <0.01 0.65 1,739

Backend WLS or EM application seems to be down Retry

goto em12c and get:

image

Backend WLS or EM application seems to be down
Retry

Wow, this bugs me.

I want to log into EM12C to check something (I love EM12C), and I get this.

I’m sure that is this going to be a rabbit hole.

ssh to the linux host, no diskspace left – great!

Really, I should disable archive log on this database, as you don’t need it (hint to solution).

Start here for some breathing room

/u01/app/oracle/admin/OEMREP/adump

Delete a shed load of files

Then run du -a . | sort -n -r | head -n 10 from /u01 to find the top 10 folders by size

This shows top 10 folders by size, great… archive log

 

Okay, let’s delete them.  RMAN to the rescue

 

oracle@machinename ~] rman

 

RMAN> connect target sys@oemrep ;

RMAN>  delete archivelog all;

 

$OMS_HOME/bin/emctl status oms

$OMS_HOME/bin/emctl stop oms -all

$OMS_HOME/bin/emctl status oms

$OMS_HOME/bin/emctl start oms

 

[oracle@musashi ~]$ $OMS_HOME/bin/emctl start oms
Oracle Enterprise Manager Cloud Control 12c Release 5
Copyright (c) 1996, 2015 Oracle Corporation.  All rights reserved.
Starting Oracle Management Server...
Starting WebTier...
WebTier Successfully Started
Oracle Management Server Successfully Started
Oracle Management Server is Up
[oracle@musashi ~]$

 

df –k

 

Filesystem     1K-blocks     Used Available Use% Mounted on
/dev/sda1      206292664 89324808 106465764  46% /
tmpfs            8085832      112   8085720   1% /dev/shm
/dev/sdc1       15349744    40672  14506308   1% /tmp

 

image

Tuesday, 7 February 2017

More ERP Analytics–how fast should my ERP be?

You know by now that we are able to patch your tools release so that you can understand how people are using your ERP.  You know you can slice and dice the data.  You know what screen is being used the most and what screen people are spending most time on!  You know this?

image

You can see from the above that I’ve shown for the last month for this client, the top applications (by page load), but also cross referencing that with page load time and also time on page – very different metrics.

So I can see the applications that are “busy”, but I can also see the applications that people spend most of their time on.  I can compare this on a monthly, weekly or daily basis.  I can also see what users are running the applications and time busy times of day.

 

Let’s look at productivity from another perspective:

Average page load time compares between two sites

clip_image001

and

clip_image001[4]

So we are comparing the average JDE page load time over an entire month across two different sites.  One has an average of .72 seconds (using data from over 313,000 individual page loads), the other has an average of 1.21 seconds for 281,000 page load samples).  Wow, is that significant?  I think so.

Do you think that asking ALL your users to wait on average another .5 of a second is appropriate?  NO WAY?  This is .49 of a second on average.  Now, the slower and smaller site loaded 544,588 pages in the month…  so if I could makes things quicker… on average, then I could get .49x544,588/3600 hours back.

74 hours of waiting around a month!  this is a pure lost productivity number – does not take into consideration user frustration.

Great, but what if you were worried about a particular application, then this could be measured also:

image

If you did make some changes, you could see how the performance has been affected, but also the time on page.