SAPGUI connection refused Windows clients HP-UX 11v1 SAP server

[Rudi van den Putte]

We're running SAP central DB and instance on one HP-UX 11v1 server. After a system copy (hardware change) the DB and instance is up and running but SAPGUI refuses to connect: "SAEconnect refused"
I already checked /etc/services: all necessary entries are included.
When using telnet with the SAPGUI port: 3200 there is also an error coming up connect on port 3200 failed, which I can't explain. telent is allowed in inetd.conf.
Any idea what is missing? Missing processes, conf - file entries mising, etc?

[Rasheed Tamton]

Check whether port 3200 is listening:

netstat -an|grep -i list|grep 32

Check any other processes are holding 3200
netstat -an|grep 3200
lsof -i:3200

[Steven E. Protter]

Shalom,

How is SAP licensed? Perhaps this product has detected the system change and needs to have its license authorization set up.

Check the connect log, wherever that is for a specific message.

SEP

[Eric SAUBIGNAC]

Bonsoir,

SEP got it. SAP is licensed based on a hardware key.

You can ask for a new license through "SAP support portal", "keys and requests"

To see existing license on the host : saplicense -show

To see hardware key : saplicense -get

Eric

[Rudi van den Putte]

Thanks for your help, but I think the saplicense is not the reason, because the SAP System has been restored on exactly the same hardware and saplicense -get says: ......
userlimit: 0 productid: R3_ORA
*** license useable ***

I also checked the netstat -an | grep 3200 result> udp 0 0 127.0.0.1.49172 127.0.0.1.3200
and further:
netstat -an|grep -i list|grep 32
with no response !!

So it seems that there is no service listening on port 3200.

What can I do? I've checked almost everything,
which service should be listening on port 3200 and how to check and set up?

THX in advance

[Rasheed Tamton]

If you have correctly added the entry in the /etc/services and opened the port 3200 in firewall. It should be ok.

[Eric SAUBIGNAC]

Bonjour,

well I am a newbie to SAP, but are you sure that dispatcher is running ? What about dispatcher log : /usr/sap//DVEBMGS00/work/dev_disp ? What about "ps -ef | grep dw\." ?

Port 3200, yes, but if instance number is 00. What is yours ?

Eric

[Rudi van den Putte]

Thanks Rasheed, thanks Eric,

I checked /etc/services sapdp00 is ok. Instance number is 00.

The dispatcher seems to be the error:
ps -ef | grep dw gave no response!
I checked the log file dev_disp, and -bingo!- it says: What can I do?

Tue Jun 24 00:08:55 2008
relno 4640patchlevel 0 atchno 1817intno 0id 21796
***LOG Q00=> DpSapEnvInit, DPStart (00 21796) [dpxxdisp.c 915]
MtxInit: -2 0 0
*** ERROR => DpSysACreate: ShmCreate (CREATE) [dpxxtool2.c 482]
*** ERROR => dispatcher already running ???? [dpxxtool2.c 483]
*** ERROR => I better EXIT before I do any (more) damage [dpxxtool2.c 484]

[Eric SAUBIGNAC]

Have you another instance of SAP running on this same system with same sap instance number 00 ?

If not you probably have some ipc still allocated. I would do a complete stop of SAP (stopsap with no arg), then complete the process with "cleanipc 00 remove". Under root login, check ipc with "ipcs -mb" and "ipcs -sb".

Eric

[Rudi van den Putte]

Thanks a lot Eric,
I think this is the right trace...
Find result of "ipcs /mb" and "ipcs /sb" below.

root@stcuxp01:../root #ipcs -mb
IPC status from /dev/kmem as of Wed Jun 25 13:33:57 2008
T ID KEY MODE OWNER GROUP SEGSZ
Shared Memory:
m 0 0x411c641e --rw-rw-rw- root root 348
m 1 0x4e0c0002 --rw-rw-rw- root root 61760
m 2 0x4120002a --rw-rw-rw- root root 8192
m 3 0x440c0373 ----------- root root 1
m 4 0x420c064b ----------- root root 1
m 1797 0x0c6629c9 --rw-r----- root root 12190424
m 6 0x06347849 --rw-rw-rw- root root 77384
m 263 0x49182df4 --rw-r--r--root root 22908
m 14088 0x00004dbe --rw-rw-rw-root sapsys 153836
m 11 0x00000000 D-rw-rw-rw-p50adm sapsys 4096
m 12 0x00000000 D-rw-r-----p50adm sapsys 423588
m 2829 0x00000000 D-rw-r----- p50adm sapsys 160000000
m 526 0x00000000 D-rw-r----- p50adm sapsys 1344

m 527 0x00000000 D-rw-r----- p50adm sapsys 1142888
m 528 0x00000000 D-rw-r----- p50adm sapsys 13059200
m 529 0x00000000 D-rw-r----- p50adm sapsys 32129
m 530 0x00000000 D-rw-r----- p50adm sapsys 3138028
m 795 0x00000000 D-rw-r----- p50adm sapsys 561488

root@stcuxp01:../root #ipcs -sb
IPC status from /dev/kmem as of Wed Jun 25 13:41:56 2008
T ID KEY MODE OWNER GROUP NSEMS
Semaphores:
s 0 0x411c641e --ra-ra-ra- root root 1
s 1 0x4e0c0002 --ra-ra-ra- root root 2
s 2 0x4120002a --ra-ra-ra- root root 2
s 3 0x00446f6e --ra-r--r-- root root 1
s 4 0x00446f6d --ra-r--r-- root root 1
s 5 0x01090522 --ra-r--r-- root root 1
s 6 0x612067a7 --ra-ra-ra- root root 1
s 7 0x732067a7 --ra-ra-ra- root root 1
s 8 0x702067a7 --ra-ra-ra- root root 1
s 9 0x692067a7 --ra-ra-ra- root root 1
s 10 0x752067a7 --ra-ra-ra- root root 1
s 11 0x632067a7 --ra-ra-ra- root root 1
s 12 0x642067a7 --ra-ra-ra- root root 1
s 13 0x662067a7 --ra-ra-ra- root root 1
s 14 0x6c2067a7 --ra-ra-ra- root root 1
s 15 0x6d2067a7 --ra-ra-ra- root root 1
s 16 0x6f2067a7 --ra-ra-ra- root root 1
s 17 0x49182df4 --ra-r--r-- root root 1
s 18 0x410c0673 --ra-ra-ra- root root 1
s 19 0x522067a7 --ra-ra-ra- root root 1

[Rudi van den Putte]

Eric,

I did zour recommendations cleanipc, etc. restarted SAP and got new errors in the dev_disp logfile. Seems to be shared memory too small issue(s).

Could you check the attached dev_disp file and let me know, how to increase the shared memory 40?

I really appreciate your help! ThX

[Eric SAUBIGNAC]

Rudi,

those shared memory claim that some SAP process are still running :

m 11 0x00000000 D-rw-rw-rw-p50adm sapsys 4096
m 12 0x00000000 D-rw-r-----p50adm sapsys 423588
m 2829 0x00000000 D-rw-r----- p50adm sapsys 160000000
m 526 0x00000000 D-rw-r----- p50adm sapsys 1344
m 527 0x00000000 D-rw-r----- p50adm sapsys 1142888
m 528 0x00000000 D-rw-r----- p50adm sapsys 13059200
m 529 0x00000000 D-rw-r----- p50adm sapsys 32129
m 530 0x00000000 D-rw-r----- p50adm sapsys 3138028
m 795 0x00000000 D-rw-r----- p50adm sapsys 561488

Note the letter 'D' at the beginning of mode field. It means that the shared memory is deleted but a process is still attached. Normally "stopsap" + "cleanipc" should have leaved a clean situation.

So :

- from p50adm user "saposcol -k"
- from orap50 user "lsnrctl stop" or "lsnrctl stop LISTENER_P50" depending the name you gave to the listener.
- from root "ps -f -u p50adm" then kill the processes shown.

Now ipcs -mb should be clean.

If it is OK restart SAP :

- from orap50 "lsnrctl start" or "lsnrctl start LISTENER_P50"
- from p50adm "startsap"

Eric

[Eric SAUBIGNAC]

Oups, did not read your previous post. I think you must have a clean situation with shared memory segments before any attempt to restart sap. And maybe, the best way to get it is to reboot the server

[Rudi van den Putte]

Eric,

all processes killed successfully, SAP restarted, dev_disp logfile says (extract):

MtxInit: -2 0 0
MBUF state OFF
MBUF opmode USE
EmInit: MmSetImplementation( 2 ).
client 0 initializing ....
InitFreeList
block size is 4096 kByte.
Using implementation std
Info: use shared memory Info: use mprotect
EsIUnamFileMapInit: shmget(IPC_PRIVATE,1677721600,IPC_CREAT | 0777) fai
errno = 12(Not enough space)
EsStdInit: unable to allocate 1600 MB
EsStdInit: try to allocate 1280 MB
EsIUnamFileMapInit: shmget(IPC_PRIVATE,1342177280,IPC_CREAT | 0777) fai
errno = 12(Not enough space)
EsStdInit: unable to allocate 1280 MB
EsStdInit: try to allocate 1024 MB
EsIUnamFileMapInit: ES base = 0xc0000000297ee000
EsStdInit: 1024 MB successfully allocated
EsIUnamFileMapInit: ES base = 0xc0000000697ee000
EsStdInit: 576 MB successfully allocated
EsStdInit: Extended Memory 1600 MB allocated
399 blocks reserved for free list.
ES initialized.

Wed Jun 25 14:36:27 2008
*** ERROR => DpWpCreate: fork (12: Not enough space) [dpuxdisp.c 709]
***LOG Q0I=> DpWpCreate, fork ( fork 12) [dpuxdisp.c 710]
*** ERROR => DpWpDynCreate: DpWpCreate [dpuxdisp.c 272]
*** ERROR => DpProcCreate: DpWpDynCreate [dpxxdisp.c 8643]
*** DP_FATAL_ERROR => DpSapEnvInit: DpProcCreate
*** DISPATCHER EMERGENCY SHUTDOWN ***
increase tracelevel of WPs
killing W0-1179 (SIGUSR2)
killing W1-1181 (SIGUSR2)
killing W2-1183 (SIGUSR2)
killing W3-1185 (SIGUSR2)
killing W4-1197 (SIGUSR2)
killing W5-1198 (SIGUSR2)
killing W6-1199 (SIGUSR2)
*** ERROR => DpWpKill: wrong pid (-1) [dpxxtool.c 2577]
*** ERROR => DpWpKill: wrong pid (-1) [dpxxtool.c 2577]
*** ERROR => DpWpKill: wrong pid (-1) [dpxxtool.c 2577]
*** ERROR => DpWpKill: wrong pid (-1) [dpxxtool.c 2577]
*** ERROR => DpWpKill: wrong pid (-1) [dpxxtool.c 2577]
*** ERROR => DpWpKill: wrong pid (-1) [dpxxtool.c 2577]
*** ERROR => DpWpKill: wrong pid (-1) [dpxxtool.c 2577]
*** ERROR => DpWpKill: wrong pid (-1) [dpxxtool.c 2577]
*** ERROR => DpWpKill: wrong pid (-1) [dpxxtool.c 2577]
*** ERROR => DpWpKill: wrong pid (-1) [dpxxtool.c 2577]
*** ERROR => DpWpKill: wrong pid (-1) [dpxxtool.c 2577]
*** ERROR => DpWpKill: wrong pid (-1) [dpxxtool.c 2577]
*** ERROR => DpWpKill: wrong pid (-1) [dpxxtool.c 2577]
*** ERROR => DpWpKill: wrong pid (-1) [dpxxtool.c 2577]
NiWait: sleep (10000 msecs) ...
NiISelect: timeout 10000 ms
Wed Jun 25 14:36:37 2008
NiISelect: TIMEOUT occured(10000 msecs)
dump system status
EsIUnamFileMapInit: ES base = 0xc0000000297ee000
EsStdInit: 1024 MB successfully allocated
EsIUnamFileMapInit: ES base = 0xc0000000697ee000
EsStdInit: 576 MB successfully allocated
EsStdInit: Extended Memory 1600 MB allocated
399 blocks reserved for free list.
ES initialized.

Wed Jun 25 14:36:27 2008
*** ERROR => DpWpCreate: fork (12: Not enough space) [dpuxdisp.c 709]
***LOG Q0I=> DpWpCreate, fork ( fork 12) [dpuxdisp.c 710]
*** ERROR => DpWpDynCreate: DpWpCreate [dpuxdisp.c 272]
*** ERROR => DpProcCreate: DpWpDynCreate [dpxxdisp.c 8643]
*** DP_FATAL_ERROR => DpSapEnvInit: DpProcCreate
*** DISPATCHER EMERGENCY SHUTDOWN ***
increase tracelevel of WPs
killing W0-1179 (SIGUSR2)
killing W1-1181 (SIGUSR2)
killing W2-1183 (SIGUSR2)
killing W3-1185 (SIGUSR2)
killing W4-1197 (SIGUSR2)
killing W5-1198 (SIGUSR2)
killing W6-1199 (SIGUSR2)
*** ERROR => DpWpKill: wrong pid (-1) [dpxxtool.c 2577]

[Rudi van den Putte]

Eric,
I've got your response now and shutting down the server currently.

I increased the shared memory pool 40 in the SAP profile, is that correct or unnecessary?

When restarting SAP, same procedure with starting the ora listener first and then SAP?

THX

[Eric SAUBIGNAC]

"When restarting SAP, same procedure with starting the ora listener first and then SAP?
" --> yes, sure.

"I increased the shared memory pool 40 in the SAP profile, is that correct or unnecessary?" --> as previously said, I am a newbie to SAP. So I don't know if this choice is OK.

But why would you want to change this parameter ? As far as I know you make this kind of adjustement at the initial configuration of SAP, or if you have some troubleshooting to do. As far I can understand you just did a hardware change, and this instance of SAP was working right before ... no ?

I see a lot of errno = 12(not enough space) in this log. What kind of changes have you really done ? Could you check HP-UX kernel parameters ? Are they still tuned for SAP ? What about swap space ?

Eric

[Rudi van den Putte]

Eric, All:

after shutdown/reboot and restart of oralistener and SAP the dev_disp file still reports ..."not enough space.." shared memory.

I already extended the size of shared pool 40 in the DVEBMGS00 profile but the message still occurs in dev_disp logfile.

Do I have to increase size on HP-UX with sam, too or what do I have to do? Please help, I think its closed to be solved.

Find the dev_disp error messages below:

THX in advance

extract of current dev_disp logfile:
--------------------------------------
InitFreeList
block size is 4096 kByte.
Using implementation std
Info: use shared memory Info: use mprotect
EsIUnamFileMapInit: shmget(IPC_PRIVATE,1677721600,IPC_CREAT | 0777) failed.
errno = 12(Not enough space)
EsStdInit: unable to allocate 1600 MB
EsStdInit: try to allocate 1280 MB
EsIUnamFileMapInit: ES base = 0xc0000000297ee000
EsStdInit: 1280 MB successfully allocated
EsIUnamFileMapInit: ES base = 0xc0000000797ee000
EsStdInit: 320 MB successfully allocated
EsStdInit: Extended Memory 1600 MB allocated
399 blocks reserved for free list.
ES initialized.

Wed Jun 25 15:38:52 2008
*** ERROR => DpWpCreate: fork (12: Not enough space) [dpuxdisp.c 709]
***LOG Q0I=> DpWpCreate, fork ( fork 12) [dpuxdisp.c 710]
*** ERROR => DpWpDynCreate: DpWpCreate [dpuxdisp.c 272]
*** ERROR => DpProcCreate: DpWpDynCreate [dpxxdisp.c 8643]
*** DP_FATAL_ERROR => DpSapEnvInit: DpProcCreate
*** DISPATCHER EMERGENCY SHUTDOWN ***
increase tracelevel of WPs

[Eric SAUBIGNAC]

shmget with "not enough space" error --> check HP-UX kernel parameters (/stand/vmunix)

[Eric SAUBIGNAC]

oups !!! check /stand/system. vmunix is the kernel itself :-(

[Rudi van den Putte]

Many thanks to All, especially Eric Saubignac, I appreciated your help. Eric, your hints finally helped to solve the problem: The root cause was - as you somehow assumed: Not enough swap space (2 GB before, now 5 GB). Which prevents the SAP dispatcher process from starting.

My last question is now: How can I increase the mounted swap space automatically from 2 GB to 5 GB at system boot? Using SAM?

Thanks a lot again.

[Rasheed Tamton]

Hi,

You need to add a new swap device of 3GB.
Can you list the output of
swapinfo -tam

[Rasheed Tamton]

>Not enough swap space (2 GB before, now 5 GB).

How come now 5GB.

Did you add any new swap device.
If you have added new swap device, then it should be added in /etc/fstab

To configure additional swap device you have to add an entry in fstab as below (in this case lvol7)

/dev/vg01/lvol7 ... swap pri=1 0 2

[Safarali]

Hi it looks like Kernel related problem
you need update the patches on your system, you need to contact your SAP team to get latest patches
We have same problems with other system and solved with patches update

Regards
Safar

[Safarali]

Hi

Also look the kernel variable with your sap documentations

Regards
Safar

[V. Nyga]

Hi,

the problem of increasing the swap is:
you have 3 logical volumes which have to be continously one after the other:
/stand, swap, / - swap is the second lvol.
So to increase swap you have to set up the server new with ignite - well, ok, I've read one tipp here in itrc how to do it in single user mode, but it's very dangerous and difficult.
So if SAP can use it, a second swap would be the (second) best solution.

HTH
Volkmar