cancel
Showing results for 
Search instead for 
Did you mean: 

system startup failed after kernel upgrade

Former Member
0 Kudos

Hi,

my kernel upgrade has failed.

the dev_disp gives the following error

      • 1889] => ESSHM: ====================> ERROR <==================== [esuxshm.c

      • ERROR => ESSHM: Error during Execution for Work Process 0 in SHM_EsCreateSha

red [esuxshm.c 1890]

      • ERROR => ESSHM: esSHMAllocateBlock returned 10 [esuxshm.c 1891]

      • 1892] => ESSHM: ====================> ERROR <==================== [esuxshm.c

      • ERROR => EgInit: EsCreateShared() failed. info='EG_GLOBAL_AREA', blocks=4003

, rc=1 [egxx.c 300]

      • ERROR => DpEmInit: EmInit (1) [dpxxdisp.c 9634]

      • ERROR => DpMemInit: DpEmInit (-1) [dpxxdisp.c 9561]

Pls help

Regards,

Ritu

Accepted Solutions (0)

Answers (5)

Answers (5)

Former Member
0 Kudos

Hello all,

I've faced the same problem after parameter optimization and in my case I've reduced the value of em/global_area_MB from 4800 to 2880. The system is up again.


em/global_area_MB + abap/shared_object_size_MB < ES/SHM_SEG_SIZE

Regards,

Serhat Yüceant

Former Member
0 Kudos

check Note 841944 - EgInit: EsCreateShared() failed. info='EG_GLOBAL_AREA'...

Reason and Prerequisites

The startup with the 6.40 kernel fails because the EG memory requirement with 640 has increased. In kernel 6.20, the requirement for EG was determined by the em/global_area_MB parameter alone. (The default is 96 MB).

With the 6.40 kernel, the ABAP Shared Object are added again and require additional address space in the EG. The size for shared objects is determined by the abap/shared_object_size_MB parameter (the default is 20 MB).

If em/global_area_MB is set too high, then the entire EG exceeds 256 MB. That is, the following applies in particular for AIX:

em/global_area_MB + abap/shared_object_size_MB < ES/SHM_SEG_SIZE

this is still the case in 700.

Former Member
0 Kudos

Hi.

Recently i to got the same error in test server.

The problem is that its because of shared memory.

So set the abap bufferzise parametr value in instance profile as half as now.

That means if the current size of the parameter is abap/buffersize = 200000 means

reduce it to 100000 now and restart the system.

Then start your instance it will surely work for you.

Regards

Vijay

JPReyes
Active Contributor
0 Kudos

seems like problems with the shared memory....

open a command prompt -> go to the kernel directory -> execute command "cleanipc XX" where XX is the system number -> then try starting the system again

Regards

Juan

Former Member
0 Kudos

Hi

Post details of dev_w0.

Rahul