Direct NFS problem in 11g
- From: Dan Norris <dannorris@xxxxxxxxxxxxx>
- To: Oracle L <oracle-l@xxxxxxxxxxxxx>
- Date: Thu, 23 Aug 2007 08:08:52 -0700 (PDT)
I have an unsupported environment running:
CentOS 4.4 in VMWare Server 1.x (2 of these)
RAC 11.1.0.6.0 (production release, not beta)
OpenFiler 2.2 serving NFS (running in a 3rd VM)
These VMs are all running on the same server (it has 16Gb of RAM and plenty of
CPU) and they're all running fine.
So, many have done this and it is working just fine. I'm trying to play with
the Direct NFS features and I know a few of you have tried this as well.
I've created an $OH/dbs/oranfstab with:
server: openfiler1
path: 172.16.163.200
export: /mnt/danrac/of1/all1 mount: /mnt/all1
server: openfiler1
path: 172.16.163.200
export: /mnt/danrac/of2/all2 mount: /mnt/all2
and they're mounted fine:
Filesystem 1K-blocks Used Available Use% Mounted on
172.16.163.200:/mnt/danrac/of1/all1
10321216 256768 9540160 3% /mnt/all1
172.16.163.200:/mnt/danrac/of2/all2
5039616 141248 4642368 3% /mnt/all2
I also changed the libodm11.so pointer to reference the NFS library per the
instructions.
I created a tablespace using this new filesystem by placing a file on /mnt/all1
and now I see this message in the alert logs of each instance:
Direct NFS: mount complete dir /mnt/danrac/of1/all1 on openfiler1 path
172.16.163.200 mntport 741 nfsport 2049
Direct NFS: mount barrier to server openfiler1 failed
Have any of you seen this before or have any suggestions? Obviously, Metalink
isn't quite ready to return search results for this new feature.
Dan
--
//www.freelists.org/webpage/oracle-l
Other related posts:
