Originally posted by: broberts
All,
I am having a severe problem with several LPARs with a like configuration. When we have an Oracle RAC cluster executing an RMAN backup to a NFS filesystem (mounted to a 3rd host which is not a member of the cluster), the LPAR that is running the RMAN backup will stop responding sometime between 30 minutes to 3 hours after the backup has started.
Our NFS mount options are set with combehind and numclust=128 as recommended by Oracle and IBM. All kernel parameters were set to IBM and Oracle recommendations regarding page_steal_method, lru settings, and the like. These settings were working fine on 5.3 TL8 SP5, but are now failing on 5.3 TL11 SP2.
We have had little luck with Oracle or IBM support at the moment (Sev1 tickets open with both). Has anyone in the community seen this issue as well?
#AIX-Forum