NIM thread blocked

Following errpt entry was logged while start up/shutdown of
cluster without any acual error.
.
LABEL        : TS_NIM_ERROR_STUCK_
Resource Name: topsvcs
Description  : NIM thread blocked
.
Probable Causes
A thread in Topology Services NIM process was blocked
Topology Services NIM process cannot get timely access to CPU
.
Detail Data
DETECTING MODULE
rsct,nim_control.C,1.19,5046
Interface name  :  en1
.
 According to a dev., It's known behavior of nim. But user hope
supress of such ambiguous error logging.

Problem summary 

In HACMP/ES, when a node is acquiring or releasing service
addresses (including at cluster startup and shutdown), it
is possible for AIX error log entries such as the
following to appear:

LABEL:          TS_NIM_ERROR_STUCK_
IDENTIFIER:     864D2CE3

Date/Time:       Wed Jan 29 23:47:20 EST

 ...

Thread which was blocked
main thread
Interval in seconds during which process was blocked
         7
Interface name
tr1


The error log entry is harmless; it is caused by
one of RSCT Topology Services's Network Interface Module
(NIM) processes taking a few seconds to process the
operations involved during IP Address Takeover. Had the
error log entry appeared in steady state operation, this
would have indicated a real thread blockage problem.

The problem was caused by the code in RSCT Topology
Services's Network Interface Module (NIM) not recognizing
a special case during IP Address Takeover (IPAT): the NIM
process's main thread takes a few seconds to complete a
certain operation. This apparent "blockage" causes the NIM
to issue (via the Topology Services daemon) the
TS_NIM_ERROR_STUCK_  AIX error log entry.

The code in the NIM is being fixed to recognize the special
IPAT case, so that no "false" TS_NIM_ERROR_STUCK_ error
entry is issued.

Comments