olsnodes make CPU spikes

After I read on otn forums.
They told they see high cpu usage, when use "olsnodes" command.

$ time olsnodes
node01
node02

real 0m0.153s
user 0m0.082s
sys 0m0.068s

And then ... solve this by delete * files at ORA_CRS_HOME/log/hostname/client/ path.

$ rm -f $ORA_CRS_HOME/log/node01/client/*

$ time olsnodes
node01
node02

real 0m0.037s
user 0m0.017s
sys 0m0.014s

$ ls $ORA_CRS_HOME/log/node01/client/*
css.log

$ time olsnodes
node01
node02

real 0m0.045s
user 0m0.016s
sys 0m0.015s

$ ls $ORA_CRS_HOME/log/node01/client/*
css1.log css.log

What about idea?

When use "olsnodes" command successfully, it scans $CRS_HOME/log/hostname/client PATH and generates a new cssxxx.log file.

metalink:729349.1 ...said it's a bug:

Olsnodes produces CPU spikes with many css*.log files in ORA_CRS_HOME/log/nodename/client directory

fixed in 11.1.0.7 and 10.2.0.4 CRS bundle Patch #2 onwards. Bundle Patch information is available in Note 756671.1

Try to check on 11.1.0.7 :

$ crsctl query crs softwareversion
Oracle Clusterware version on node [test01] is [11.1.0.7.0]

$ ls $ORA_CRS_HOME/log/test01/client/ | wc -l
32738

$ time olsnodes
test01
test02

real 0m0.030s
user 0m0.015s
sys 0m0.014s

OK... No problem on 11.1.0.7