Prepare: Shadow instance not starting and RFC error

[ThomasKjelsrud]

Hi all

I am running PREPARE because we want to upgrade from R/3 4.7 Ext.set 2.00 to ERP 2005 SR2 on Windows 2003 and MSSQL 2005.

I have a problem when trying to start the Shadow Instance that is created during prepare (I try using the SAPMMC, "./startsap DVEBMGS<NR> R3" command). The Message Server starts, but the Dispatcher is not starting.

In the dev_disp log file for the shadow instance I find this:

Wed Jan 03 18:05:25 2007

• Error 11 while initializing OS dependent part.

• ERROR => DpEmInit: EmInit (1) [dpxxdisp.c 9544]

• ERROR => DpMemInit: DpEmInit (-1) [dpxxdisp.c 9473]

• DP_FATAL_ERROR => DpSapEnvInit: DpMemInit

• DISPATCHER EMERGENCY SHUTDOWN ***

increase tracelevel of WPs

NiWait: sleep (10000ms) ...

NiISelect: timeout 10000ms

NiISelect: maximum fd=1

NiISelect: read-mask is NULL

NiISelect: write-mask is NULL

Wed Jan 03 18:05:35 2007

NiISelect: TIMEOUT occured (10000ms)

dump system status

Any ideas on how to proceed??

However, when I try to continue prepare I get this error message in the PSUPGRFC.LOG file:

1 ETQ201XEntering upgrade-phase "RUN_RSUPGRFC" ("20070103171730")

2 ETQ367 Connect variables are set for standard instance access

4 ETQ399 System-nr = '01', GwService = 'sapgw01'

4 ETQ399 Environment variables:

4 ETQ399 dbs_mss_schema=t01

4 ETQ399 auth_shadow_upgrade=0

4 ETQ010 Date & Time: 20070103171730

4 ETQ265 Starting report "RSUPGRFC" with variant "SAP_SHDREM" in batch

4 ETQ359 RFC Login to: System="T01", Nr="01", GwHost="vmkomplett", GwService="sapgw01"

2 ETQ231 RFC Login failed

2EETQ360 RFC of "subst_start_report_in_batch" failed:

2EETQ361 code/exception : open

2EETQ362 key : RFC_ERROR_COMMUNICATION

2EETQ364 message :

2EETQ399 Connect to SAP gateway failed

2EETQ399 Connect_PM DEST=T01, GWHOST=vmkomplett, GWSERV=sapgw01, SYSNR=01

2EETQ399

2EETQ399 LOCATION CPIC (TCP/IP) on local host

2EETQ399 ERROR partner 'vmkomplett:3301' not reached

2EETQ399

2EETQ399 TIME Wed Jan 03 17:17:31 2007

2EETQ399 RELEASE 700

2EETQ399 COMPONENT NI (network interface)

2EETQ399 VERSION 38

2EETQ399 RC -10

2EETQ399 MODULE nixxi.cpp

2EETQ399 LINE 2764

2EETQ399 DETAIL NiPConnect2

2EETQ399 SYSTEM CALL connect

2EETQ399 ERRNO 10061

2EETQ399 ERRNO TEXT WSAECONNREFUSED: Connection refused

2EETQ399 COUNTER 1

2EETQ399

4 ETQ359 RFC Login to: System="T01", Nr="01", GwHost="vmkomplett", GwService="sapgw01"

2 ETQ231 RFC Login failed

2EETQ236 RFC Login failed with return code -1

1EETQ204 Upgrade phase "RUN_RSUPGRFC" aborted with severe errors ("20070103173337")

Return code -1 indicates an error with client, username or password for the RFC connection. When I look at the RFC in SM59 in my system, the hosts that the R/3 connection "SAP_UPGRADE_SHADOW_SYSTEM" points to in sapp01r3, which is totally unknown to me. Our host is called something else.

I have no idea on where to start looking from here.

Help is greatly appreciated and needed.

Regards,

Thomas

[Former Member · Accepted solution]

Hi Thomas,

To reset the upgrade to initial state I proceed as followed:

1. Stop the upgrade server. If you can't do this from Upgrade Assistant, kill the java process.

2. Delete put directory

3. Start prepare.bat again (don't forget to create put directory)

4. After prepare.bat has finished, start upgrade server as normal (java -cp ua.jar UaServer)

5. Try to loggon on with upgrade assistant

6. If this works stop upgrade server (i normally kill the java process) and replace SAPup, R3trans, tp in ../usr/sap/put/bin

7. Restart upgrade server and connect with upgrade assistant

If you change any files while upgrade server is running you have to restart them!

ATTENTION!!!

You have to use tp and R3trans from Kernel for WebAS 7.00 because you are upgrading to ERP2005 -> this is your TARGET RELEASE. Don't change any other tp, R3trans in other directories, only if upgrade assistant told you to do so. Upgrade assistant comes with an right version of the source release kernel!!

Regards,

Daniel

[ThomasKjelsrud · Accepted solution]

Hi all,

I have reset Prepare and followed the instructions given by Daniel above. So far so good, but I am still waiting to reach the Shadow Instance start up. I will keep this thread open until I have been able to get the shadow instance running.

I greatly appreciate the support! Thanks a lot!

Regards,

Thomas

[Former Member · Accepted solution]

Hi Thomas,

going thru the communication b/n u and the rest of the friends, i have i small query. what phase are you, exactly in? since, shadow instance will be only up just before, SPDD activity and will be downafter wards (at the time of upgrade).

but we cann't connect to shadow instance, at the time of prepare phase. at the time of prepare phase we will just give the settings of shadow instance, number like instance number, ports etc.

we can connect to shadow instance only after this <b>"START_SHDI_FIRST"</b> phase.

please go thru the upgrade document. or else give me your id, i will mail u the docs.

with Br,

Rajesh,.

[ThomasKjelsrud · Accepted solution]

HI Rajesh,

I was in the Installation phase of PREPARE, in a "sub-phase" called RUN_RSUPGRFC. I get an error as you have seen that made me think that there was an issue with the RFC connection to the shadow instance, and since the shadiw instance was not starting I assumed that this was the reason.

Do you have experience that can tell me is the shadow instance is supposed to start during the prepare phase? I do not mean up in terms of being able to log in normally, but in order to get RFC connection to the shadow instance..

Regards,

Thomas

Message was edited by:

Thomas Kjelsrud

[Former Member · Accepted solution]

hi thomas,

as you have re-started the prepare from the first.. i hope you have cleared tht etc/service file, removed the previous settings.

as per my understanding, i guess you didn't start the actual upgrade. if tht is so, then it is too early to worry abt tht shadow instance. R3up will bring the shadow instance only at the system downtime ( for more on this refer upgrade manual).

that time, the central instance will be down and shadow instance will be up, once we complete our SPDD activity, then shadow instance will down complete forever.

i had forwarded my documentation on upgrade to your mail id.

with best rgds,

Rajesh.

<i>award suitable points</i>

[ThomasKjelsrud]

Hi all!

I would like to give all of you 10 points for helping me solve this issue, but unfortunately I can't. Thanks to everyone helping me, and I hope that I can be of help to you in the future.

Best Regards,

Thomas Kjelsrud

[former_member185954]

Hi Thomas,

Glad that the problem is resolved, its not about the points, its about learning, I hope you too would contribute in this forum of with the experience you are gaining.

Regards,

Siddhesh

[Former Member]

Hi Thomas,

Its really good to hear, that your problem is resolved. ALL the best.

BR,

Rajesh.

[hamendra_patel]

Thomas/ rajesh ,

i was reading this thread from last few days , since i am also upgrading solution manager 3.2 to 4.0 on AIX .

I also got same problem , my upgrade phase stopped with error :

"SEVERE ERROR: phase CHECKGROUP_END4 not valid."

Here is the Detail :

I started upgrade phase and get error :

SERVE ERROR : "1311" is occupied in file services for the send daemon talk protocol "

So i started Prepare Initialization phase again and changed the Shadow host from 11 to 12 and also changed port no's for all required ports :

SEND DAEMON LISTEN PORT = 7712

> COLLECT DAEMON TALK PORT = 7412

> SEND DAEMON TALK PORT = 7312

> COLLECT DAEMON LISTEN PORT = 7912 .

After this change when i tried to restart Upgrade it failed soon , with

error :

>> 15:41:43 UPGRADE: START OF PHASE SHDINST_CHK1

Severe internal inconsistency, aborted.

Pls sugest what to do ?

Thanks

Hamendra Patel

[Former Member]

Hi Hamendra

<i> i suggest you to create a new msg in this forum. since it will be viewed the rest of the users and they will come up with there suggestions too</i>.

comming to your problem,

i guess its the problem of ports..it should be, something like this.

1. SHD INSTANCE NUMBER = 03

2. SHD send d talk port = 1303

3. SHD send d listen port = 1203

4. SHD collect d listen port = 3903

5. SHD collect d talk port = 4003

6. SHD GW SERVICE = sapgw03

based upon the shadow instance number, provide the appropriate values. pls, check that ports are not used by any other services. check the values in R3up.log file.

i hope this info helps you!

with regards,

raj.

<i>pls award suitable points.</i>

[Former Member]

Hi,

You've have to use tp and R3trans of your target release! Kernel 83 is the latest one. Use this one. I've used this version this week for an upgrade.

Regards

Daniel

[ThomasKjelsrud]

Daniel,

LATEST:

When I try to reset prepare using the "SAPup reset prepare" from usr\sap\put\bin I get a message saying:

"The upgrade has already been started. PREPARE cannot be reset". But I haven't started the upgrade! Any ideas??

Sorry about this, but I can't seem to find the R3trans and tp that I am looking for. Can you please advise on the path to download it?

I guess I only need these two files, or should I do an entire kernel upgrade? I am currently on patch level 154 on WAS 6.40.

Regards,

Thomas

Message was edited by:

Thomas Kjelsrud

[Former Member]

Hi Thomas,

Please double check that your upgrade is running in PREPARE phase. You can get the latest tp and R3trans ( i guess 162 is the latest ) from http://service.sap.com/swdc --> Download --> support packs and packages --> Additional components --> SAP kernel --> SAP Kernel 64 bit --> Select your OS --> Database Independet

Thanks

Prince

[ThomasKjelsrud]

Hi Prince,

When I restart the Upgrade Server and Upgrade Gui, a message at the bottom is telling me that "SAPup started in PREPARE mode". This should indicate that I am in fact in PREPARE? I also look at the CHECKS.LOG, and it is telling me that I am in the Installation phase. So why is it telling me that the upgrade has started, and how can I reset PREPARE in order to start over?

Ok, so I only need those two files (R3trans and tp), and not the entire kernel?

Thanks!

Regards,

Thomas

[Former Member]

Hi Thomas,

Yes, I recommend to start PREPARE from begining (including delete put directory) and also delete the service of the shadow instance (in the regestry or with sapstartsrv.exe).

Use netstat -a to see what ports are used on your host.

Regards

Daniel

[Former Member]

Hi Thomas,

It is always a good Idea to use the latest SAPup Version. Download the latest version

from marketplace and put it /usr/sap/put/bin (don't delete the old one, rename it).

I also recommend to start prepare from beginning with this new SAPup version (Don't forget to restart upgrade Server first!!!).

Is the shadow instance installed one the same host as the R/3 System?

Also check if there is anything running that blocks or uses port 3201, 3301, 3601.

regards,

daniel

[ThomasKjelsrud]

Hi Daniel,

I will upgrade SAPup now, and see what happends.

I assume you want me to do a complete restart of prepare (including deleting the put directory)?

The shadow system is installed on the same host as the R/3 system.

How can I check that nothing is running on the ports? Using netstat command?

Thanks!

Regards,

Thomas

[former_member185954]

Hi Thomas,

please also ensure that you are using latest versions of R3trans and tp tools.

Regards,

Siddhesh

[ThomasKjelsrud]

Siddhesh,

Should I use the latest R3trans and tp for my target release or my source release?

I find these in the latest kernel?

Regards,

Thomas

[Former Member]

Hi Thomas,

Intressting probem. I had a similar problem some time ago.

Some questions: In what phase of PREPARE are you?

Have you applied the latest tp, R3trans, SAPup, Corrections for Upgrade, after you first start the Upgrade Server (this is a very importent information, because if the versions are not correct you will always run in trouble, believe me, at the moment I make an upgrage from 4.6C to ERP2005 / MSSQL2005).

And don't think about R3up. You make an upgrade to to WebAS 7.00 so you only have to use SAPup 7.00 (R3up is used until WebAS 6.40).

Regards

Daniel

[ThomasKjelsrud]

Hi Daniel,

My biggest concern is that the shadow instance is not starting. In the process list for the shadow instance, the message server is running, but the dispatcher starts ok and then stops after about a minute. Do you have any idea of what might cause this, and possibly how to fix it?

I am in the Installation phase of Prepare, in a phase called "RUN_RSUPGRFC".

I have checked the CHECKS.LOG file, and it tells me that no upgrade of R3trans, tp, disp+work is needed. There is no information regarding SAPup. I have applied the latest corrections for the upgrade "ERP2005SR2_8-20000914.SAR" and unpacked. The files FIX_ERP2005SR2.UPG and FIX_ERP2005SR2_8.info are in the usr\sap\put directory.

Please let me know if you think you have any information that might help us get through this.

Regards,

Thomas

Message was edited by:

Thomas Kjelsrud

[Former Member]

Hi Thomas,

please look into ‘STARTSFI.LOG’ & ‘DEVTRACE.LOG’ for the error details. and try remote log on, using the RFC connection,

"SAP_UPGRADE_SHADOW_SYSTEM". check whtr instance number is correctly mentioned &

check whtr the following actions are done already.

• You have to login as root and enter: the shadow dispatcher instance sapdp<shadow instance number> in /etc/services with number 32<insance number>

• Please create the user for the shadow instance by executing

SQL script crshdusr.sql. [which is available in “/DEPOT” directory].

try to create a copy of START_DVEBMGS<instance no>_<hostname> with name START_DVEBMGS<instance_no>.

and try starting it manually.

<u><b>Stopping Shadow instance</b></u>

cd /usr/sap/put

cd bin

R3up stopshd

Enter the path of your upgrade directory [/usr/sap/put]:

Upgrade directory = /usr/sap/put

<u><b>Starting Shadow instance</b></u>

R3up startshd

Enter the path of your upgrade directory [/usr/sap/put]:

Upgrade directory = /usr/sap/put

with br,

Raj.

[ThomasKjelsrud]

Hi Raj,

I have checked the services file, and everything seems to be correct. It looks like this:

sapmsSHDT01 3601/tcp #Upgrade shadow instance

sapdp01 3201/tcp #Upgrade shadow instance

sapgw01 3301/tcp #Upgrade shadow instance

When I try to execute the crshdusr.sql script that you mention (it was located in the \put\bin directory...) I get a few error messages:

Msg 195, Level 15, State 5, Line 1

'feedback' is not a recognized SET option.

Msg 195, Level 15, State 5, Line 2

'termout' is not a recognized SET option.

Msg 195, Level 15, State 5, Line 4

'serverout' is not a recognized SET option.

Msg 155, Level 15, State 2, Line 15

'number' is not a recognized CURSOR option.

Msg 102, Level 15, State 1, Line 23

Incorrect syntax near 'R3USR'.

Msg 102, Level 15, State 1, Line 29

Incorrect syntax near ':'.

Msg 102, Level 15, State 1, Line 39

Incorrect syntax near 'cur'.

Msg 102, Level 15, State 1, Line 43

Incorrect syntax near 'cur'.

Msg 156, Level 15, State 1, Line 55

Incorrect syntax near the keyword 'THEN'.

Msg 102, Level 15, State 1, Line 61

Incorrect syntax near 'cur'.

Msg 102, Level 15, State 1, Line 68

Incorrect syntax near 'cur'.

Msg 195, Level 15, State 5, Line 81

'termout' is not a recognized SET option.

Msg 102, Level 15, State 1, Line 84

Incorrect syntax near 'dbms_output'.

Msg 102, Level 15, State 1, Line 91

Incorrect syntax near ';'.

So I am guessing that this script is not the one I am after.

You ask me to try remote log in, using RFC connection

"SAP_UPGRADE_SHADOW_SYSTEM". Can you please tell me how you want me to perform this?

When I try to start the shadow instance using startsap (R3up does not exist in usr\sap
put\bin) I get an error message saying that it failed. SAPT01-log shows me this:

"startsap name=T01 nr=01 SAPDIAHOST=vmkomplett -wait"

Any more ideas?

Thanks a lot for you efforts!

Regards,

Thomas

[Former Member]

Hi Thomas,

i feel, there is some problem with the startsap Script or RFC_connection itself.

open the script file, and check with is the profile it searching for while starting SAP.

re-check RFC setting by loggin using 'DDIC'.

open this parameter file "SHADOW.TP0" and cross check all the parameters.

regarding that RFC.

1)trying logging into the system using client='000', user-id="ddic" and the password.

2)execute t-code 'sm59' and open R/3 connections.

3)open RFC "SAP_UPGRADE_SHADOW_SYSTEM" & check the RFC settings.

4)click on "test connection" & after success, click on "remote logon".

in case if you are using Clustering concept, you may face this issue. so try specifying the correct hostname(if possible IPaddress).

hope this will reslove your issue.

with Br,

Raj.

<i> award suitable points </i>

[ThomasKjelsrud]

Hi Raj,

Just to verify.. you are asking me to log into the source instance (system number 00), right? the shadow system is not starting, and is not possible to log in to.

Also, I assume that the RFC setting in SM59 in source instance (00) should be pointing to the host where the shadow instance is running, with the system number of the shadow instance (01)? Is there anything else that should be in the "SAP_UPGRADE_SHADOW_SYSTEM" RFC connection? Such as gateway host and gateway service? This would the "hostname" and sapgw01 (gateway for the shadow instance). Correct?

Thanks.

Regards,

Thomas

[Former Member]

Dear Thomas,

yes you are right. you have to log into instance '00' only, using 'ddic'. R3UP will automatically update the necessary info, like instance number & hostname at the time of prepare phase only.

check wthr you have posted an entry in "/etc/services/host" file like

"IPADDRESS HOSTNAME" .

if you are facing issue with host name, then try specifing with IPaddress of the Server in RFC. for more this check with your network adminstrator, if you r not able to ping to tht server from your system.

changing the settings of RFC and try again.

even we too faced the same problem, but when i tried to log in using RFC, it worked, so we did out SPDD using RFC only.

hope this will reslove the issue.

with Best regards,

Raj.

<i> award suitable points for my reply</i>

[ThomasKjelsrud]

Hi Raj, thanks for your quick replies!

I do however, wonder if you have any ideas as to why my shadow instance is not starting. I am sure that I will be able to get the RFC connection running, if I could only get the shadow instance to start.

I checked the SHADOW.TP0 file:

#####################################################################

1. This is the global parameter file for the different transport

2. parameters.

3. -------------------------------------------------------------------

4. First we specify global default values for each parameter,

5. later the system specific incarnation of special parameters

#####################################################################

#

daylight_shutdown = yes

clientcascade = true

with_tacob = no

lock_eu = yes

impdp_by_event = yes

#

repeatonerror = 8

stoponerror = 9

kernelmove = false

###touchall = no (IS NO LONGER USED)

#

#

1. The next option must be set to care for adjusted customer modifications,

2. because the mass activator would use the old timestamp, if the structure

3. in inacticve/active nametab are the same - but: SAP-object must be generated

4. due to customer adjustments =>> reset_timestamps = yes for MVNTABS

reset_timestamps = yes

#

#

nbufform = 1

tp_version = 264

To be honest, I didn't get much useful out of this.. maybe you can?

Regards,

Thomas

Message was edited by:

Thomas Kjelsrud

[former_member185954]

Hi Thomas,

I checked your first post again.

>>When I look at the RFC in SM59 in my system, the hosts that the R/3 connection "SAP_UPGRADE_SHADOW_SYSTEM" points to in sapp01r3, which is totally unknown to me. Our host is called something else.

make the host point to your correct hostname, however as RAJA said login as DDIC in client 000 and make the correction and then try.

Regards,

Siddhesh

[ThomasKjelsrud]

Hi Siddhesh,

Thanks for your efforts!

Even if I could make the changes in the RFC settings correct, I doubt that PREPARE would be able to continue, and my shadow system is not running! I get an error in SAPMMC that the Dispatcher i stopped. The Message Server is running though. I think this might be the main reason why the RFC is failing.

You can have a look at the log files in one of my earlier post in this thread.

Regards,

Thomas

[Former Member]

Hi Thomas,

i was under the impression tht, their might be some problem while specifying inputs, so i asked you to cross check this Shadow.TP0 file. sometime it happens, tht SAP will take wrong inputs, even you give the correct [it will pick from previous installations].

even i am unable to trace out anything from this file.

Raj.

[former_member185954]

Hi,

Also check the following SAP Notes:

Note 525677 - Problems when starting the shadow instance

Note 398100 - Shadow instance during upgrade

Regards,

Siddhesh

[former_member185954]

Hi thomas,

I think manual stopstart of shadow instance should be done with : R3up startshd

also , I see that the gateway service is not present.

Check your services file and see if sapgw01 exists and points to 3301

Regards,

Siddhesh

[ThomasKjelsrud]

Hi Siddesh,

From where should I try to run R3up startshd?

My put directory is C:\usr\sap\put..

I have also checked that sapgw01 is in the services file and it points to 3301.

Can I start the sapgw01 manually somehow?

Thank you for your help!

Regards,

Thomas

[former_member185954]

Hi thomas,

Manual start procedure is given in the notes i specified. Here are some more points to look at eliminate errors caused by them.

Make sure your c:\usr\put directory doesn't have any access issues.

Make sure these ports exists in the services file

36<Instance number Shadow instance> ...Message Server Port

33<Instance number Shadow instance> ...Gateway Port

32<Instance number Shadow instance> ...Dispatcher Port

Enter the message server port of the shadow instance as follows in the %WINDIR%\system32\drivers\etc\services file:

sapmsSHD<SAPSID> 36<instance number of the shadow instance>/tcp

With target release 6.40, additional libraries are required for the SAP kernel. To install these, execute the R3dllins.exe program from the NT\I386\NTPATCH directory on the Upgrade Master CD.

As of Release 7.00, the R3dllins.exe program is in the UM<U/N>_WINDOWS_<Platform>\DBINDEP\NTPATCH directory on the Upgrade Master DVD.

Is there any other dev_w* file in the <DIR_PUT>/<SID>/DVEBMGS*/work directory, if yes, it should contain more information regarding the error, if there is such file, post it here in this forum.

Regards,

Siddhesh

[ThomasKjelsrud]

Hi again Siddhesh,

I am looking in the <DIR_PUT>/<SID>/DVEBMGS/work directory for <i>any</i> dev_w files, but there are none. No dev_w* files whatsoever...

The services file looks like this with regard to shadow instance ports:

sapmsSHDT01 3601/tcp #Upgrade shadow instance

sapdp01 3601/tcp #Upgrade shadow instance

sapgw01 3601/tcp #Upgrade shadow instance

I ran the R3dllins.exe before I started prepare, so I don' think this is the reason.

In light of this information, do you have any more ideas?

Thank you!

Regards,

Thomas

Message was edited by:

Thomas Kjelsrud

[former_member185954]

Hi Thomas,

What is the instance number of your original source system ?

Is it 00 , cause if its 01 then the Shadow instance should have instance number 02.

Also your services file doesn't seem to be correct,

instead of this...

sapmsSHDT01 3601/tcp #Upgrade shadow instance

sapdp01 3601/tcp #Upgrade shadow instance

sapgw01 3601/tcp #Upgrade shadow instance

it should be this...

sapmsSHDT01 3601/tcp #Upgrade shadow instance

sapdp01 3201/tcp #Upgrade shadow instance

sapgw01 3301/tcp #Upgrade shadow instance

All this assuming that your source system has instance number 00 and shadow instance is having instance number 01.

Regards,

Siddhesh