Deployment Server - XE update 1

Spike

Active Member
I'm new to OneWorld and I tried performing my first update. I tried installing the XE update 1 last night and had some performance problems. I selected our 4 environments for it to update and checked the backup option. The update has been running 14 hours now, and it looks like it is still backing up the first environment. Does anyone know why this backup is taking so long? I haven't even got to the "build packages" process yet. Our Deployment server is a PIII 800 MHZ single processor. Is our Deployment server undersized or is something else causing the problem? I checked the hard drive and I can tell the size of the backup for this environment is increasing in size, so it looks like its working, only slower than anticipated. None of the other environments have a backup file, so I can tell it hasn't started on them yet.

Ryan
 
what is going on here is that the "backup" is creating a package in the planner directory of all your existing objects.

Since Update 1 touches so many objects, believe me this is going to take a while. My bet is that you wanted to back up and if so, you're going to have to wait a while. If it takes 6 hours to build a package and your doing this on 4 pathcodes - 6x4 = 24 hours.

Let it go. It'll work itself out.

Larry - how does this affect live production ? Is this scaremongering ? There is no effect on your production environment by doing this since it is only copying your specifications to a local file on the deployment server.

Jon Steel
Xe Upgrade Specialist
ESfU specialist



ERP Sourcing
http://www.erpsourcing.com
[email protected]
 
maybe I was scaremongering. My apologies Ryan. Its still not something I would do given the success rate of JDE's major updates.

Jon - question for you: Don't ASUs update Central Objects in the database (via Spec merge)? If so, then wouldn't a failed merge on some of his objects leave Ryan without any place to restore/copy those objects from?

Repentantly,

Larry Jones
[email protected]
OneWorld B733.1, SP 11.3
HPUX 11, Oracle SE 8.1.6
SandBox: OneWorld XE
 
yes. Central objects certainly get updated with both ASU's and ESU's. However, you must remember that Central Objects do not get run against - Central Objects are there so that the package builds can create "Runtime" objects which are run against.

For example, if I deleted F98741 - even though all of the development ER are located in there - I will not affect production users even if I did this in the production pathcode. Those users interpret the event rules from the GBRSPEC.DDB file either locally, on the Citrix Server or on the Enterprise Server. The GBRSPEC.DDB is created from the F98741 as a snapshot.

Hopefully, Ryan performed a database backup before he attempted the update - therefore if he does crash out Central Objects, he can restore them without affecting anyone.

I am in the process of writing a whitepaper on a process we use to implement ESU's in a heavy development organization - check this space !

Jon Steel
Xe Upgrade Specialist
ESfU Specialist

ERP Sourcing
http://www.erpsourcing.com
[email protected]
 
I did the XE Update 1 on a Dual 450mhz with 512mb ram deployment server on 1
environment and selected the backup option and it probably took about 8
hours.
 
yes AND NO (one small qualification). Please feel to correct me if I am
wrong here.

Central objects also gets hit anytime someone modifies a version and
overrides the template specs and checks in the version. If those specs do
not exist on another client in any of GBRSPEC, RDASPEC, and RDATEXT spec
files, JITI will occur. If the Central Objects code is missing in F98740,
F98740, F98741, F98760, F98761 and F98762, the version is corrupt and won't
JITI. If it won't JITI, it won't run at RUNTIME.



----- Original Message -----
From: "altquark" <[email protected]>
To: <[email protected]>
Sent: Tuesday, March 20, 2001 3:24 PM
Subject: Re: Deployment Server - XE update 1


> yes. Central objects certainly get updated with both ASU's and ESU's.
However, you must remember that Central Objects do not get run against -
Central Objects are there so that the package builds can create "Runtime"
objects which are run against.
>
> For example, if I deleted F98741 - even though all of the development ER
are located in there - I will not affect production users even if I did this
in the production pathcode. Those users interpret the event rules from the
GBRSPEC.DDB file either locally, on the Citrix Server or on the Enterprise
Server. The GBRSPEC.DDB is created from the F98741 as a snapshot.
>
> Hopefully, Ryan performed a database backup before he attempted the
update - therefore if he does crash out Central Objects, he can restore them
without affecting anyone.
>
> I am in the process of writing a whitepaper on a process we use to
implement ESU's in a heavy development organization - check this space !
>
> Jon Steel
> Xe Upgrade Specialist
> ESfU Specialist
>
> ERP Sourcing
> http://www.erpsourcing.com
> [email protected]
> --------------------------
> Visit the forum to view this thread at:
>
http://198.144.193.139/cgi-bin/wwwthreads/showflat.pl?Cat=&Board=OW&Number=7
603
> + - - - - - - - - - - - - - - - - - - - - - - - - - - - - - +
> This is the JDEList One World / XE Mailing List.
> Archives and information on how to SUBSCRIBE, and
> UNSUBSCRIBE can be found at http://www.JDELIST.com
> + - - - - - - - - - - - - - - - - - - - - - - - - - - - - - +
>
>



JD Nowell
OW: B7332
ES: AS400 V4R4 CO: DB2/400 SP: 11.2
Users: 250 TSE Users: 100
 
Ryan,

Someone I know also said that they selected the "backup" option and that it took
an absurd amount of time to run.. His explanation to me was that the update
affects virtually every object in the system and thus must back them all up.

I do not think it is the size of your deployment server that is the issue, but
the size of the update :)

I hope this info helps :)

t








[email protected] on 03/20/2001 02:22:52 PM

Please respond to [email protected]








To: [email protected]

cc: (bcc: Tim Wozniak/Flint)



Subject: Deployment Server - XE update 1








Subject: Deployment Server - XE update 1

I'm new to OneWorld and I tried performing my first
update. I tried installing
the XE update 1 last night and had some performance
problems. I selected our 4
environments for it to update and checked the backup
option. The update has
been running 14 hours now, and it looks like it is still
backing up the first
environment. Does anyone know why this backup is taking
so long? I haven't
even got to the "build packages" process yet. Our
Deployment server is a PIII
800 MHZ single processor. Is our Deployment server
undersized or is something
else causing the problem? I checked the hard drive and
I can tell the size of
the backup for this environment is increasing in size,
so it looks like its
working, only slower than anticipated. None of the
other environments have a
backup file, so I can tell it hasn't started on them
yet.

Ryan


--------------------------
Visit the forum to view this thread at:
http://198.144.193.139/cgi-bin/wwwthreads/showflat.pl?Ca
t=&Board=OW&Number=7579

----- End forwarded message -----
+ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - +
This is the JDEList One World / XE Mailing List.
Archives and information on how to SUBSCRIBE, and
UNSUBSCRIBE can be found at http://www.JDELIST.com
+ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - +
 
Jon,

yes, Central ojects don't usually get hit by production users (see other poster's note on this).

However to me its a big deal to have to restore all or part of your database (some Oracle sites keep everything in one database). When you need to restore an object or two. Thats like using a D9 caterpillar to make your sand castle with. Central Objects is still data and it changes on our site every day.

My original concern that I voiced so poorly was that anytime you apply a "fix" you should do so initially in a non-production environment, test it, re-apply customizations if needed, and then copy to production ...
(In the process of writing the above statement I just realized how poorly we are served by JDE when they distribute these mega ESU's to us ...)

The other concern I have is a deep distrust of the Spec Merge process based on the experience that it can corrupt objects to the point that we have needed to copy from pristine / JD7333.

My 2 cents,

Larry Jones
[email protected]
OneWorld B733.1, SP 11.3
HPUX 11, Oracle SE 8.1.6
SandBox: OneWorld XE
 
Hi Ryan,
I just installed Update 1 couple of days ago. The backup is what takes the
most time. It took up about 10 hours to do the back up and we have a PIII
single processor machine too. But you have selected all 4 environment so
it will take a long time to backup all those. I would have done just
development or Prototype to start with.

Anyway if you have any specific questions you can send me an email directly
at [email protected].

----- Original Message -----
From: <[email protected]>
To: <[email protected]>
Sent: Tuesday, March 20, 2001 4:22 PM
Subject: Deployment Server - XE update 1


> Subject: Deployment Server - XE update 1
>
> I'm new to OneWorld and I tried performing my first
> update. I tried installing
> the XE update 1 last night and had some performance
> problems. I selected our 4
> environments for it to update and checked the backup
> option. The update has
> been running 14 hours now, and it looks like it is still
> backing up the first
> environment. Does anyone know why this backup is taking
> so long? I haven't
> even got to the "build packages" process yet. Our
> Deployment server is a PIII
> 800 MHZ single processor. Is our Deployment server
> undersized or is something
> else causing the problem? I checked the hard drive and
> I can tell the size of
> the backup for this environment is increasing in size,
> so it looks like its
> working, only slower than anticipated. None of the
> other environments have a
> backup file, so I can tell it hasn't started on them
> yet.
>
> Ryan
>
>
> --------------------------
> Visit the forum to view this thread at:
> http://198.144.193.139/cgi-bin/wwwthreads/showflat.pl?Ca
> t=&Board=OW&Number=7579
>
> ----- End forwarded message -----
> + - - - - - - - - - - - - - - - - - - - - - - - - - - - - - +
> This is the JDEList One World / XE Mailing List.
> Archives and information on how to SUBSCRIBE, and
> UNSUBSCRIBE can be found at http://www.JDELIST.com
> + - - - - - - - - - - - - - - - - - - - - - - - - - - - - - +
>
>
 
I appreciate all the help people have provided. I (with the help of JDE) finally found out why the backup was taking so long. The JDE.INI on our deployment server had the JDE logging set to FILE. This was creating a massive JDE.LOG file on our deployment server. We set the logging to NONE and tried the update again. I appears to be backuping up environments much quicker now. Luckily, we are not live yet and can learn from our mistakes.

Ryan
 
I encountered the same problem. I then aborted the run and deselected the backup option and re-run the program in attended mode. I first took a backup of the deployment server.
 
Gopal,

how long did it take to install the update with no backup?

Larry Jones
[email protected]
OneWorld B733.1, SP 11.3
HPUX 11, Oracle SE 8.1.6
SandBox: OneWorld XE
 
I remember sitting at the DS for the duration of the upgrade, it took less than 3 hours attended for PY7333.

>>> [email protected] 22/03/01 07:11:46 >>>
Gopal,

how long did it take to install the update with no backup?

Larry Jones
[email protected]
OneWorld B733.1, SP 11.3
HPUX 11, Oracle SE 8.1.6
SandBox: OneWorld XE
--------------------------
Visit the forum to view this thread at:
http://198.144.193.139/cgi-bin/wwwthreads/showflat.pl?Cat=&Board=OW&Number=7751
+ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - +
This is the JDEList One World / XE Mailing List.
Archives and information on how to SUBSCRIBE, and
UNSUBSCRIBE can be found at http://www.JDELIST.com
+ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - +
 
Back
Top