#1131 IMM: duplicate immnd process was observed starting opensaf with PBE.

[Hrishikesh]

changeset : 5697
Setup : 4 nodes cluster with SLES 64bit.

Problem:

Opensaf is configured with single PBE and 100K objects loaded.
Upon starting opensaf, 100K objects were loaded but PBE process didnot start.
It was also observed that there are two immnd processes spawned and running

Syslog clearly says that "PBE is starting" but it never started successfully and no error messages observed.

Sep 22 19:33:15 SLES-SLOT1 osafimmnd[32038]: NO STARTING PBE process.
Sep 22 19:33:15 SLES-SLOT1 osafimmnd[32038]: NO pbe-db-file-path:/home/PBE/imm.db VETERAN:0 B:0
Sep 22 19:33:15 SLES-SLOT1 osafimmnd[32038]: NO Implementer connected: 3 (safClmService) <15, 2010f>
Sep 22 19:33:15 SLES-SLOT1 osafimmnd[32038]: NO implementer for class 'SaClmNode' is safClmService => class extent is safe.
Sep 22 19:33:15 SLES-SLOT1 osafimmnd[32038]: NO implementer for class 'SaClmCluster' is safClmService => class extent is safe.
Sep 22 19:33:15 SLES-SLOT1 osafclmna[32097]: Started
Sep 22 19:33:15 SLES-SLOT1 osafclmna[32097]: NO safNode=SC-1,safCluster=myClmCluster Joined cluster, nodeid=2010f
Sep 22 19:33:15 SLES-SLOT1 osafamfd[32106]: Started
=========

[Neelakanta Reddy]

1. The write permission is not given for mounted shared filesystem where PBE resides.

2. immnd forked with PBE process

Sep 22 19:33:15.014270 osafimmnd [32038:immnd_proc.c:1506] >> immnd_forkPbe
Sep 22 19:33:15.014276 osafimmnd [32038:immnd_proc.c:1519] TR exec-pbe-file-path:/usr/lib64/opensaf/osafimmpbed
Sep 22 19:33:15.014284 osafimmnd [32038:immnd_proc.c:1542] NO pbe-db-file-path:/home/PBE/imm.db VETERAN:0 B:0
Sep 22 19:33:15.014414 osafimmnd [32038:osaf_secutil.c:0065] >> handle_new_connection
Sep 22 19:33:15.015334 osafimmnd [32038:mds_c_api.c:1605] TR svc UP process_info EXIST, svc:26 cnt:2, adest:2010fbe536019
Sep 22 19:33:15.015323 osafimmnd [32038:immnd_proc.c:1584] T5 Parent osafimmnd, successfully forked /home/PBE/imm.db, pid:32091

1. But the ps output is still showing the PBE as immnd library path (may be struck at fork because of permission issue)

opensaf 32038 1 1 19:32 ? 00:00:19 /usr/lib64/opensaf/osafimmnd --tracemask=0xffffffff
opensaf 32059 1 0 19:33 ? 00:00:00 /usr/lib64/opensaf/osaflogd
opensaf 32072 1 0 19:33 ? 00:00:00 /usr/lib64/opensaf/osafntfd
opensaf 32079 32072 0 19:33 ? 00:00:00 /usr/lib64/opensaf/osafntfimcnd active
opensaf 32086 1 0 19:33 ? 00:00:00 /usr/lib64/opensaf/osafclmd
opensaf 32091 32038 0 19:33 ? 00:00:00 /usr/lib64/opensaf/osafimmnd --tracemask=0xffffffff

1. When rerun-again with write permission to mounted filesystem, the problem is not seen.

[Hans Feldt]

Duplicate of #1077
Why are you testing such old version as 5697 of opensaf?
RC2 was released yesterday and should be the focus in test.

[Hans Feldt]

1077 was not fixed in 4.3 and 4.4 though

[Anders Bjornerstedt]

Was the problem resolved ?

[Neelakanta Reddy]

when the cluster stopped and started again the problem is not seen

[Anders Bjornerstedt]

So do we close the ticket or was there a problem with insufficiently clear
error indication ?

[Neelakanta Reddy]

As Hans, pointed it may be duplicate of #1077.
After having the changeset of #1077, still the problem is seen, please re-open the ticket.