In the previous post I described how to cope with issues before start applying the bundlepatch. Now this post is about an issue I run into after applying this bundlepatch.
During execution of “rootcrs.pl -postpatch” there were errors starting several resources. All of these issues were caused by the network resource which could not be started after the patch. The output, or clusterware alert.log, showed something like this:
2015-12-15 11:24:10.375 [ORAROOTAGENT(7672)]CRS-5017: The resource action "ora.net1.network start" encountered the following error: 2015-12-15 11:24:10.375+CRS-5008: Invalid attribute value: <Network name> for the network interface. For details refer to "(:CLSN00107:)" in "D:\app\oracle\diag\crs\<hostname>\crs\trace\crsd_orarootagent_system.trc".
Nobody changed the network interface names, no one modified the clusters network configuration. So what?
After a couple of hours of rebooting, investigating and tearing hair I stumbled upon a MOS note Windows: gpnpd fails to start as “oifcfg iflist -p -n” does not show any network interface (Doc ID 1967213.1) which included the final hint. In my case the GPNPD was up by the way. But the GI was no longer able to read the network configuration:
D:\>oifcfg iflist D:\>
Not a single interface available. As stated in the MOS note, the “wimgmt” service was not running. I disabled that service because it was using some of GI’s files and thus I could not run “opatch apply -local”. My mistake…. Re-enabling the service fixed this issue. Everything runs fine again.