disp+work.EXE (Dispatcher) - Stopped

[joo_migueldimas]

Hi gurus,

I´m facing with an error when I try to start the SAP System in SAPMMC. The disp+work.EXE (dispatcher) is stopped with gray semaphore!

Can you help me to solve this error.. please..

I checked the log *dev_disp:*

...

NiBufReceive starting

MsINiRead: received 114 bytes

MSG received, len 110+4, flag 3, from MSG_SERVER , typ 0, key -

Received 4 bytes from MSG_SERVER

Received opcode MS_NOOP from msg_server, reply MSOP_OK

MsOpReceive: ok

MsSendKeepalive : keepalive sent to message server

NiIRead: hdl 3 recv would block (errno=EAGAIN)

Sat Oct 24 21:18:51 2009

NiIPeek: peek for hdl 3 timed out (r; 1000ms)

NiIRead: read for hdl 3 timed out (1000ms)

DpHalt: no more messages from the message server

DpHalt: sync with message server o.k.

detach from message server

***LOG Q0M=> DpMsDetach, ms_detach ()

NiBufSend starting

NiIWrite: hdl 3 sent data (wrt=110,pac=1,MESG_IO)

MsINiWrite: sent 110 bytes

MsIDetach: send logout to msg_server

MsIDetach: call exit function

DpMsShutdownHook called

NiBufISelUpdate: new MODE -- (r-) for hdl 3 in set0

SiSelNSet: set events of sock 1512 to: ---

NiBufISelRemove: remove hdl 3 from set0

SiSelNRemove: removed sock 1512 (pos=3)

SiSelNRemove: removed sock 1512

NiSelIRemove: removed hdl 3

MBUF state OFF

AdGetSelfIdentRecord: > <

AdCvtRecToExt: opcode 60 (AD_SELFIDENT), ser 0, ex 0, errno 0

AdCvtRecToExt: opcode 40 (AD_MSBUF), ser 0, ex 0, errno 0

AdCvtRecToExt: opcode 40 (AD_MSBUF), ser 0, ex 0, errno 0

blks_in_queue/wp_ca_blk_no/wp_max_no = 1/300/10

LOCK WP ca_blk 1

make DISP owner of wp_ca_blk 1

DpSemRq: key: 2, units: 1, timeout: -1

DpSemRel: key: 2, units: 1

DpRqPutIntoQueue: put request into queue (reqtype 1, prio LOW, rq_id 15)

MBUF component DOWN

NiICloseHandle: shutdown and close hdl 3 / sock 1512

NiBufIClose: clear extension for hdl 3

MsIDetach: detach MS-system

cleanup EM

EsCleanup ....

EmCleanup() -> 0

Es2Cleanup: Cleanup ES2

***LOG Q05=> DpHalt, DPStop ( 1880)

Good Bye .....

[Former Member]

Hi

Pl go to the command prompt and run the "R3trans -d". It will create a file as "trans.log". Pl open & check the same you will know the reason of dispatcher stopped..

Regards

Vimal

[joo_migueldimas]

Hello Vimal,

I already did that command on prompt. At first time that gave me an error 0012, and after having done one more time the kernel upgrade this error disappear and now that string gave the output/result 000.

You can see in the following thread the message that I wrote two days ago about this issue "r3trans -d":

But unfortunately the disp+work.exe it still didn´t start!! Can you give other/any tip what I have to do... I tried to analyze some log files like dev_disp ; dev_w0 ; dev_dispatcher ... but I don´t know yet how I solve this!!

Best regards,

João Dimas - Portugal

[Former Member]

What changes did you make before the error was observed the first time? Did you perform a kernel update oder anything else before the problem occured the first time?

In case of a kernel update:

You may check the dev_* logfiles for a file version mismatch.

Make sure you close all applications which may access files in your kernel directory.

Starting with release 7.x at system startup sapcpe is used to copy files from the kernel directory to instance_dir/exe. Check if you got old files there.

Regards

[joo_migueldimas]

Hi Christian Kaunzner,

Yes I did a Kernel Upgrade to solution manager system (700) with two stacks (abap & java) in windows server 2003 X32 with oracle database and this system have two parts, this is, a unicode (uc) & non-unicode (nuc) section.

I already checked some log files, among them, those dev_w0 (Question: Which dev_w(X) Do I must check?! I check only the dev_w0... this is enough or I have to check others dev_w(X)?

There is no other open application!!

I don't get to find that instance_dir/exe, where I can see that? Can you tell me please?

Best regards,

João Dimas - Portugal

[Former Member]

Hi João,

dev_w0 should be enough. A version mismatch should be listed at the first one or two pages.

The instance directory is located at /usr/sap// DVEBMGSxx/exe (which is used to start the system). You can verify this by checking the start profile. You will find something like this (see below) iif sapcpe is used:

#----

-

---

1. Copy SAP Executables

#----

-

---

Start_Program_00 = immediate $(DIR_CT_RUN)sapcpe$(FT_EXE) pf=$(_PF)

If sapcpe is enabled you may try to:

- stop SAP system

- backup and delete contents of .../DVEBMGSxx/exe

- start SAP system again.

This way you are sure not to use old files when starting the system.

What do you mean by "this system have two parts, this is, a unicode (uc) & non-unicode (nuc) section." ?

Is your ABAP system non-unicode?

Regards

Christian

[joo_migueldimas]

Hi Christian,

When I open the start profile START_DVEBMGS01_SLMQLSMP1 in <Drive>:\usr\sap\<SID>\SYS\profile I see this:

....

************************

SAPSYSTEMNAME = SMP

SAPGLOBALHOST = SLMQLSMP1

SAPSYSTEM = 01

INSTANCE_NAME = DVEBMGS01

DIR_CT_RUN = $(DIR_EXE_ROOT)\$(OS_UNICODE)\NTI386

DIR_EXECUTABLE = $(DIR_INSTANCE)\exe

DIR_PROFILE = $(DIR_INSTALL)\profile

PF = $(DIRPROFILE)\SMP_DVEBMGS01_SLMQLSMP1

#----

---

1. Copy SAP Executables

#----

---

Start_Program_00 = immediate $(DIR_CT_RUN)\sapcpe$(FT_EXE) pf=$(_PF)

#----

---

1. Start ABAP database

....

So if I realize this means that files could be the problem... and what I must to do is delete all content files in this directory <Drive>:\usr\sap\<SID>\DVEBMGS01\exe this is what you wanted to say? But for SAP start this isnu00B4t required?! Itu00B4s weird... I must change all files in kernel directory (<Drive>:\usr\sap\<SID>\SYS\exe\uc\NTI386) again or itu00B4s not necessary?

Answer to your question the ABAP system is Unicode!!! But this have two parts on kernel directory, the nuc and uc. And when I did the kernel upgrade I change both parts... Did I do well this?

Best regards,

Jou00E3o Dimas - Portugal

[florin_radulea]

Hi João Dimas

If your sistem is 4.7 you can try to alter database name to rename it global name.

I had same problem in the past and this was my solving.

alter database rename global_name to SID

Regards,

Florin

[joo_migueldimas]

Hi Florin Radulea,

My system is 7.00, is a solution manager system with oracle DB and windows 2003 server!

Best regards

João Dimas - Portugal