Skip to Content
0
Former Member
Mar 12, 2008 at 01:26 PM

Portal java servers shutdown and does not restart

21 Views

HI there,

We are running Netweaver 7.0 SPS12 JSDK 1.4.2_13. We have 3 Central Instances (BI, CRM, ECC) on one server, and 3 Dialog Instances (BI, CRM, ECC) on another, with the SCS and ASCS clustered on a different server. AMD64 / Windows 2003 Server / MSCS cluster.

All 3 java stack instances go down at the same time on either server inconsistently with the following error in dev_jcontrol where the java jlaunch processes do not start up again....

[Thr 9856] *** WARNING => JsfOpenShm: ShmCreate (69, 312160, CREATE) failed(rc = 4 SHM segment already exists) [jsfxxshm.c 715]

[Thr 9856] *** ERROR => JsfOpenShm: Shm version mismatch (shm "JSF-ADM-CORRUPT", code "JSF-ADM-015") [jsfxxshm.c 761]

[Thr 9856] *** ERROR => Can't create shared memory segment 69 (rc = 8) [jcntrxx.c 1759]

[Thr 9856] *** ERROR => Can't initialize JControl Administration [jcntrxx.c 273]

[Thr 9856] JControlCloseProgram: started (exitcode = -1)

[Thr 9856] *** ERROR => JsfCloseShm: FiDetachIndex(SESSION) failed (rc = 6 invalid argument) [jsfxxshm.c 1243]

[Thr 9856] *** ERROR => JsfCloseShm: FiDetachIndex(ALIAS) failed (rc = 6 invalid argument) [jsfxxshm.c 1250]

[Thr 9856] *** ERROR => JsfCloseShm: FiDetachIndex(SERVICE) failed (rc = 6 invalid argument) [jsfxxshm.c 1257]

[Thr 9856] JControlCloseProgram: good bye... (exitcode = -1)

To start up the java instances again, the jcontrol process has to be killed and the cluster restarted from SMICM.

Any ideas? Please help!

kind regards,

Clinton