RE: Oracle over NFS
- From: "Lange, Kevin G" <kevin.lange@xxxxxxxxxx>
- To: <troach@xxxxxxxxx>, "Oracle Discussion List" <oracle-l@xxxxxxxxxxxxx>
- Date: Mon, 19 Apr 2010 16:11:29 -0500
We currently have multiple Oracle 10.2.0.3 and 10.2.0.4 along with a 2
node RAC cluster working on Netapp.
Our standard connections for our RAC installation is
na7g1:/vol/rac_oradata - /mnt_na7g1_rac_oradata nfs - yes
hard,vers=3,suid,nointr,proto=tcp,rsize=32768,wsize=32786,bg,noac,timeo=
600
While or standard connections for non-rac installations are
na9g2:/vol/ttcp - /mnt_na9g2_ttcp nfs - yes
hard,vers=3,suid,intr,proto=tcp,rsize=32768,wsize=32768
For non-rac setups, we use their volume software so that each of our
databases is in its own volume and therefore do not affect any other
instance if some issue arrises.
We do not use Rman.
We have an archive script and a backup script that incorporates the
Netapp Snapshots to give us our backup protection along with Standby
databases at a seperate colocation. The archive script pushes logs and
applies to the standbys every 15 minutes. The databases undergo a hot
backup every night where the tablespaces are put into backup mode and
then the Netapp Snapshot occures. We have sucessfully used these hot
backups to recover. We also haved a cold backup happening every week
where the dbs are brought down (again, only need one down at a time
since they are in their own netapp volume) and snapshotted,
For the rac setups, we currently just use a big hammer .... i.e. we
bring down the instances and asm and take a snapshot. Very Very ugly
and old-school. Working on something beter . (but then, its only a
test instance so we are not in a hurry to change yet).
________________________________
From: oracle-l-bounce@xxxxxxxxxxxxx
[mailto:oracle-l-bounce@xxxxxxxxxxxxx] On Behalf Of Thomas Roach
Sent: Monday, April 19, 2010 1:12 PM
To: Oracle Discussion List
Subject: Oracle over NFS
Hi,
We are currently on Oracle RAC 10.2.0.4 (4 nodes) on Linux x86_64 RHEL
5.3. We are currently using block devices on Hitachi but have invested
in a NetApp. Our 2 options are NFS (which the Unix admin favors) and
Block devices over Fiber Channel. In trying the various NFS option, we
have come across 2 mount options that Oracle must use. (noac and
actimeo=0). With noac, the performance is horrible (about 10 MB/s) and
with actimeo=0 we get about 240 MB/s which is the maximum bandwidth the
interface will support (3 - 1Gbps Nics on the NetApp side). With the
actimeo=0, we get better performance but see a large amount of network
discards. He has essentially said I am putting it back to noac to get
rid of the network discards. I said if the performance goes back to 10
MB/s then we can't move to NFS over NetApp. I wanted to get some insight
from the group as I get conflicting information when researching this
issue.
Does anyone out there have Oracle RAC running over NFS to NetApp and if
so, what mount options are you using? Any specific configurations on the
NetApp side? Forgot to mention that we are using a single Cisco switch
between the NetApp and Oracle servers running with jumbo frames.
Thanks!
Tom
This e-mail, including attachments, may include confidential and/or
proprietary information, and may be used only by the person or entity
to which it is addressed. If the reader of this e-mail is not the intended
recipient or his or her authorized agent, the reader is hereby notified
that any dissemination, distribution or copying of this e-mail is
prohibited. If you have received this e-mail in error, please notify the
sender by replying to this message and delete this e-mail immediately.
Other related posts:
