SAP Kernel 700 201 on HP-UX

[Former Member]

Hi Gurus,

Anyone of you faced the below error after the kernel upgrade 700 to patch level 201.

I have upgraded in one system it worked fine.Strangely ,Now when I upgraded in the other system with the same release it is giving the below error, the disp is throwing emergency shut down..

Thu Aug 20 22:32:04 2009

• ERROR => gateway (pid 20205) died [dpxxdisp.c 16538]

DpGwCheck: appc_adm.stat = 2

• DP_FATAL_ERROR => Gateway could not be started - I better exit now

• DISPATCHER EMERGENCY SHUTDOWN ***

increase tracelevel of WPs

kill(20207,0) successful -> process alive

killing W0-20207 (SIGUSR2)

kill(20208,0) successful -> process alive

Please help.

Thanks,

Jagadish

[Former Member]

Sap provided workaround for the issue by setting up a parameter gw/resolve_phys_addr = 0

[Former Member]

Hmm, when i remember correctly I had this once. The reason was that one of the workprocesses used port 3300 (as outgoing port) to connect to the dispatcher and the gateway process was not able to bind to 3300.

After a clean restart, everything was ok. If this does not work, check the gateway trace file. If you see something like 'unable to bind port', then please verify no other process is using the gateway port. You can use the lsof command, if you have it:

 root# lsof -i tcp:3300

Use netstat if you don't have lsof, but you cannot see the process then:

root# netstat -an | grep 3300

If you see this, somebody already has the port:

tcp        0      0  *.3300                 *.*                     LISTEN

If you cannot find a port issue then, try a clean stopsap and cleanipc remove.

Best regards, Michael

[Former Member]

Hi,

Thanks for th reply!

This is what the gateway trace with the new kernel

hcgwdd1:hd6adm 27> more dev_rd.old

-

---

trc file: "dev_rd", trc level: 1, release: "700"

-

---

Fri Aug 21 00:18:25 2009

***LOG S00=> GwInitReader, gateway started ( 6523) [gwxxrd.c 1757]

systemid 274 (HP (IA-64) with HP-UX)

relno 7000

patchlevel 0

patchno 201

intno 20050900

make: single threaded, Unicode, 64 bit, optimized

pid 6523

gateway runs with dp version 242000(ext=110000) (@(#) DPLIB-INT-VERSION-242000-UC)

gateway (version=700.2009.02.03)

gw/local_addr : 0.0.0.0

CCMS: AlInitGlobals : alert/use_sema_lock = TRUE.

Fri Aug 21 00:18:26 2009

Bind service sapgw06 (socket) to port 3306

***LOG S41=> GwInitShTbl, ShmCreate ( 04) [gwxxtool.c 1525]

GwInitReader: try to attach to shared memory

GwPrintMyHostAddr: my host addresses are :

1 : [10.16.18.21] 10.16.18.21 (HOSTNAME)

2 : [127.0.0.1] localhost (LOCALHOST)

***LOG S10=> GwInitReader, GwAddrToHostCanon () [gwxxrd.c 2060]

But when I revert the old kernel it is picking up the shared memory properly

make: single threaded, Unicode, 64 bit, optimized

pid 9238

gateway runs with dp version 232000(ext=109000) (@(#) DPLIB-INT-VERSION-232000-UC)

gateway (version=700.2007.10.01)

Fri Aug 21 00:35:45 2009

gw/local_addr : 0.0.0.0

CCMS: AlInitGlobals : alert/use_sema_lock = TRUE.

CCMS: Initalizing shared memory of size 60000000 for monitoring segment.

Fri Aug 21 00:35:46 2009

Bind service sapgw06 (socket) to port 3306

***LOG S41=> GwInitShTbl, ShmCreate ( 04) [gwxxtool.c 1525]

GwInitReader: try to attach to shared memory

GwPrintMyHostAddr: my host addresses are :

1 : [10.16.18.21] 10.16.18.21 (HOSTNAME)

2 : [127.0.0.1] localhost (LOCALHOST)

Full qualified hostname =

DpSysAdmExtCreate: ABAP is active

DpSysAdmExtCreate: VMC (JAVA VM in WP) is not active

Fri Aug 21 00:35:53 2009

DpShMCreate: sizeof(wp_adm) 25168 (1480)

DpShMCreate: sizeof(tm_adm) 14328608 (28600)

DpShMCreate: sizeof(wp_ca_adm) 28800 (96)

DpShMCreate: sizeof(appc_ca_adm) 9600 (96)

DpCommTableSize: max/headSize/ftSize/tableSize=500/16/552064/552080

DpShMCreate: sizeof(comm_adm) 552080 (1088)

DpSlockTableSize: max/headSize/ftSize/fiSize/tableSize=0/0/0/0/0

DpShMCreate: sizeof(slock_adm) 0 (104)

DpFileTableSize: max/headSize/ftSize/tableSize=0/0/0/0

DpShMCreate: sizeof(file_adm) 0 (72)

DpShMCreate: sizeof(vmc_adm) 0 (2008)

DpShMCreate: sizeof(wall_adm) (104064/88352/80/192)

DpShMCreate: sizeof(gw_adm) 48

DpShMCreate: SHM_DP_ADM_KEY (addr: 0xc00000003a000000, size: 15145680)

DpShMCreate: allocated sys_adm at 0xc00000003a000000

DpShMCreate: allocated wp_adm at 0xc00000003a0021c0

DpShMCreate: allocated tm_adm_list at 0xc00000003a008410

DpShMCreate: allocated tm_adm at 0xc00000003a008470

DpShMCreate: allocated appc_ca_adm at 0xc00000003adb9810

DpShMCreate: allocated comm_adm at 0xc00000003adbbd90

DpShMCreate: system runs without slock table

DpShMCreate: system runs without file table

DpShMCreate: allocated vmc_adm_list at 0xc00000003ae42a20

DpShMCreate: allocated gw_adm at 0xc00000003ae42aa0

DpShMCreate: system runs without vmc_adm

DpShMCreate: allocated ca_info at 0xc00000003ae42ad0

MtxInit: -2 0 0

[markus_doehr2]

> GwPrintMyHostAddr: my host addresses are :

> 1 : [10.16.18.21] 10.16.18.21 (HOSTNAME)

> 2 : [127.0.0.1] localhost (LOCALHOST)

> ***LOG S10=> GwInitReader, GwAddrToHostCanon () [gwxxrd.c 2060]

> GwPrintMyHostAddr: my host addresses are :

> 1 : [10.16.18.21] 10.16.18.21 (HOSTNAME)

> 2 : [127.0.0.1] localhost (LOCALHOST)

> Full qualified hostname =

To me this seems more a network issue - the second kernel may be more restrictive.

You have your /etc/hosts and /etc/nsswitch.conf properly configured? Seems your system has neither a hostname nor a full qualified domain name.

Markus