Error during system copy

[Former Member]

Hello Experts,

I am doing a heterogeneous copy. During the Start Instance phase of the copy the instance fails to start. The services are started, the DB is up too, everything looks fine but still copy is not continuing. Here are my findings.

Below is the response of niping -v

Hostname/Nodeaddr verification:

===============================

Hostname/Nodeaddr verification:

===============================

Hostname of local computer: newsaptest (NiMyHostName)

Lookup of hostname: newsaptest (NiHostToAddr)

--> IP-Addr.: 10.1.1.207

Lookup of IP-Addr.: 10.1.1.207 (NiAddrToHost)

--> Hostname: newsaptest

Lookup of hostname: localhost (NiHostToAddr)

--> IP-Addr.: 127.0.0.1

Lookup of IP-Addr.: 127.0.0.1 (NiAddrToHost)

--> Hostname: localhost

External ip: 127.0.0.1

Internal Ip: 10.1.1.207.

Below is the error msg from /work folder

CPICTRC7378 file says: ERROR => NiPConnect2: SiPeekPendConn failed for hdl 3 / sock 13

(SI_ECONN_REFUSE/111; I4; ST; 10.1.1.207:3306) [nixxi.cpp 2770]

• ERROR => GwIConnect: GwConnect to newsaptest / 3306 failed (rc=NIECONN_REFUSED) [gwxx_mt.c 296]

dev_w says:

ES initialized.

B db_con_shm_ini: WP_ID = 9, WP_CNT = 17, CON_ID = -1

I *** ERROR => shmat(39452703,0x(nil),SHM_RND) (12: Cannot allocate memory) [shmux.c 1597]

B dbtbxbuf: Shm Segment 19: Cannot attach

B ***LOG BBB=> ADM message TBX buffer: function shmcreate0 returns RC = 256 [dbtbxbuf#2 @ 16094] [dbtbxbuf1609 4]

B ***LOG BZL=> internal error in table buffer: table buf init fail [dbtbxbuf#2 @ 1701] [dbtbxbuf1701 ]

B dbtbxbuf: return code (sap_rc): 2, Buffer TBX_GENERIC will not be available

B db_tblinit failed

M *** ERROR => ThCallHooks: event handler db_init for event CREATE_SHM failed [thxxtool3.c 261]

M *** ERROR => ThIPCInit: hook failed [thxxhead.c 2084]

M ***LOG R19=> ThInit, ThIPCInit ( TSKH-IPC-000001) [thxxhead.c 1523]

M in_ThErrHandle: 1

M *** ERROR => ThInit: ThIPCInit (step 1, th_errno 17, action 3, level 1) [thxxhead.c 10468]

Can someone please help here.

Thanks!

Rahul.

[markus_doehr2 · Accepted solution]

> I am doing a heterogeneous copy

Just for completion: You need to have a certified migration consultant on-site to do heterogeneous migration, otherwise you'll loose SAP support for the target system (see http://service.sap.com/osdbmigration --> FAQ).

> dev_w says:

>

> ES initialized.

> B db_con_shm_ini: WP_ID = 9, WP_CNT = 17, CON_ID = -1

> I *** ERROR => shmat(39452703,0x(nil),SHM_RND) (12: Cannot allocate memory) [shmux.c 1597]

> B dbtbxbuf: Shm Segment 19: Cannot attach

> B ***LOG BBB=> ADM message TBX buffer: function shmcreate0 returns RC = 256 [dbtbxbuf#2 @ 16094] [dbtbxbuf1609 4]

> B ***LOG BZL=> internal error in table buffer: table buf init fail [dbtbxbuf#2 @ 1701] [dbtbxbuf1701 ]

> B dbtbxbuf: return code (sap_rc): 2, Buffer TBX_GENERIC will not be available

> B db_tblinit failed

This seems to be an OS specific IPC shared memory configuration issue.

What OS do you use?

Markus

[Former Member · Accepted solution]

Hi Markus,

I am the same guy who posted the question about import abap phase couple of days earlier :). anyways reg. the issue.

I am using a Linux SuSE11 version. Raised a message to SAP and waiting for their reply too. What do you think might have gone wrong here?

Thanks!

Rahul.

[markus_doehr2 · Accepted solution]

> I am the same guy who posted the question about import abap phase couple of days earlier :). anyways reg. the issue.

>

> I am using a Linux SuSE11 version. Raised a message to SAP and waiting for their reply too. What do you think might have gone wrong here?

Did you configure /etc/sysctl.conf to reflect the shared memory settings? Is the suse-sapinit package installed?

Markus

[Former Member · Accepted solution]

Hi Markus,

The package suse-sapinit is not installed on the system, nor could I find the package for SLES 11 SP1 on novell site.

WE have not even edited /etc/sysctl.conf file but we have done installation on similar kind of machine and it has worked without these 2.

Thanks,

Rahul

[markus_doehr2 · Accepted solution]

Sorry, the name for that component is now sapconf (it's installed with 'SAP application server base' during the initial setup.

The defaults may have been enough for the other systems, for this copy apparently not.

What's the content of /etc/sysctl.conf?

Markus

[Former Member · Accepted solution]

Hi Markus,

You are right. Made me realize that guides are not that bad a thing :). Found the note 1310037 specific for SuSE11 and the system admin has starting building the system from scratch.

Besides the issue, due to really think sapconf is that critical a component that would result in such weird errors? Do you advise following the note while setting up the system will get me across this error?

Thanks!

Rahul.

[markus_doehr2 · Accepted solution]

sapconf contains Linux kernel specific settings (/etc/sysctl.conf). All *nix like operating systems have kernel parameters, on Linux they can be adapted on-line, other Os'ses (e. g. HP-UX) require a kernel relink and a reboot if those are changed.

There are various memory models for a SAP system on Linux (see Note 941735 - SAP memory management for 64-bit Linux systems), depending on what memory model is used you need to adapt the kernel to deal with those. The default limits are rather low by default (for security reasons) so you have to increase them - by installing the package.

On top comes the fact, that there are user related limits (enter "ulimit -a" or "limit" in a shell), those are adapted too with that package (/etc/security/limits.conf). If those limits are too low or not increased at all the default memory allocation fails with errors you see in your trace files.

Markus

[Former Member · Accepted solution]

Looks like i have hit the jackpot then :). Starting all over again. Shall keep posted if the error goes off..

Thanks for your suggestion.

Rahul.