One node from cluster environment is out.

classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

One node from cluster environment is out.

faraz.gilane
Dear Erman Arslan,

We have 2 database servers running 1 instance of oracle db each pointing to the same ASM storage on Oracle RAC cluster environment.

Recently we have updated the Kernel on both the nodes from kernel.x86_64 3.10.0-327.18.2.el7  to kernel.x86_64 3.10.0-514.2.2.el7.

Please note we have rebooted only NODE 2 so the new kernel update has been applied only in NODE 2.
Node 1 is still using the old kernel version.

After updating NODE 2 kernel we have seen that the OracleASM package is unavailable now[Node 2 is down however Node 1 is still up and connecting to the DB]. After contacting RHEL & Oracle support we have realized that a new OracleASM module version has been released which is compatible with the new kernel version. [The older OracleASM module is not compatible with the new kernel version]

Finally we have the following queries and we need your expertise & assistance on the same.
Q1.) If we update the OracleASM module in Node 2, should we also update the same in Node 1 along with its kernel version.
Q2.) When the new OracleASM module has been updated will it require any major configuration changes.
Q3.) In case we decide not to update Node 1 now and update only Node 2 OracleASM module will it cause any issues or have any major impact.
Q4.) Your views & advise on the situation.

Current Oracle ASM module version in both the nodes.
kmod-oracleasm.x86_64            2.0.8-15.el7            @rhel-7-server-eus-rpms

Proposed new Oracle ASM module version is still not final. [Waiting for response from RHEL support]

Red Hat Enterprise Linux Server release 7.2 (Maipo) in both the nodes.

Thanks in advance
Reply | Threaded
Open this post in threaded view
|

Re: One node from cluster environment is out.

ErmanArslansOracleBlog
Administrator
Hi Faraz,

In such a scenario, the recommendation is to update/upgrade  each node in a RAC environment at the same time (not rolling).

INFO:

*Oracle Clusterware and Real Application Clusters support mixed servers in the same cluster, as long as the servers run the same operating system (e.g., same OS binary). You can build a cluster of machines of different speeds and size. In addition, Oracle supports the rolling upgrade of the operating system in a cluster when both versions of the operating system are certified with the version of Oracle Database you are running. Note mixed operating system versions are only supported for the duration of an upgrade (i.e., within 24 hours) --the cluster should never be operated with mixed operating systems for an extended period. Oracle Clusterware and Real Application Clusters do not support heterogeneous platforms (e.g, machines having different chip architectures) in the same cluster.

*ASMLib rolling upgrade will not harm anything.  RAC = partial cluster is down (only one node or so, but the system continues in operation on the surviving nodes )...that is the goal of RAC.   Also, the UEK kernel updates the ASMLib automatically.

*ASMLib is a support library for the Automatic Storage Management (ASM).  The Oracle ASMLib product has a kernel-specific component, the ASMLib driver. This is provided in an RPM package which must be updated whenever the Linux kernel is updated

*The use of a later ASMLib version under the newer kernel is essentially a product upgrade. Depending on versions used, rolling product upgrade i.e. the concurrent use of the old and new versions (such as when using RAC clusters) may not be compatible, therefore always review ASMLib upgrade documentation.

Check the following document, as it gives you the way and answer to your questions..
How To Upgrade ASMLib Kernel Driver as Part of Kernel Upgrade? (Doc ID 1391807.1)
Reply | Threaded
Open this post in threaded view
|

Re: One node from cluster environment is out.

faraz.gilane
Hi ,

Thanks for the quick response. We appreciate your support, we will keep you updated on the progress.
Reply | Threaded
Open this post in threaded view
|

Re: One node from cluster environment is out.

faraz.gilane
Dear Erman,

We updated the kmod-oracleasm package in the Node 2 server and everything worked out perfectly without any configuration changes or major issues.

As kernel updates are not major OS changes, RAC allows 2 nodes on cluster to have different kernel versions.

After a couple of days we updated Node 1 too with the latest kernel & kmod-oracleasm package.

Thanks for the support.
Reply | Threaded
Open this post in threaded view
|

Re: One node from cluster environment is out.

ErmanArslansOracleBlog
Administrator
good to know Faraz. Thanks for the feedback 

6 Şub 2017 19:50 tarihinde "faraz.gilane [via Erman Arslan's Oracle Forum]" <[hidden email]> yazdı:
Dear Erman,

We updated the kmod-oracleasm package in the Node 2 server and everything worked out perfectly without any configuration changes or major issues.

As kernel updates are not major OS changes, RAC allows 2 nodes on cluster to have different kernel versions.

After a couple of days we updated Node 1 too with the latest kernel & kmod-oracleasm package.

Thanks for the support.


If you reply to this email, your message will be added to the discussion below:
http://erman-arslan-s-oracle-forum.2340467.n4.nabble.com/One-node-from-cluster-environment-is-out-tp2040p2103.html
To start a new topic under RAC, email [hidden email]
To unsubscribe from Erman Arslan's Oracle Forum, click here.
NAML