Storagetek 6140 - chunk size? - veritas and sun cluster tuning?

hi, we've just got a 6140 and i did some raw write and read tests -> very nice box!
current config: 16 fc-disks (300gbyte / 2gbit/sec): 1x hotspare, 15x raid5 (512kibyte chunk)
3 logical volumes: vol1: 1.7tbyte, vol2: 1.7tbyte, vol3 rest (about 450gbyte)
on 2x t2000 coolthread server (32gibyte mem each)
it seems the max write perf (from my tests) is:
512kibyte chunk / 1mibyte blocksize / 32 threads
-> 230mibyte/sec (write) transfer rate
my tests:
* chunk size: 16ki / 512ki
* threads: 1/2/4/8/16/32
* blocksize (kibyte): .5/1/2/4/8/16/32/64/128/256/512/1024/2048/4096/8192/16384
did anyone out there some other tests with other chunk size?
how about tuning veritas fs and sun cluster???
veritas fs: i've read so far about write_pref_io, write_nstream...
i guess, setting them to: write_pref_io=1048576, write_nstream=32 would be the best in this scenario, right?

I've responded to your question in the following thread you started:
https://communities.oracle.com/portal/server.pt?open=514&objID=224&mode=2&threadid=570778&aggregatorResults=T578058T570778T568581T574494T565745T572292T569622T568974T568554T564860&sourceCommunityId=465&sourcePortletId=268&doPagination=true&pagedAggregatorPageNo=1&returnUrl=https%3A%2F%2Fcommunities.oracle.com%2Fportal%2Fserver.pt%3Fopen%3Dspace%26name%3DCommunityPage%26id%3D8%26cached%3Dtrue%26in_hi_userid%3D132629%26control%3DSetCommunity%26PageID%3D0%26CommunityID%3D465%26&Portlet=All%20Community%20Discussions&PrevPage=Communities-CommunityHome
Regards
Nicolas

Similar Messages

Is Veritas- or Sun Cluster needed for RAC in a Solaris Environment

Is a Veritas- or Sun Cluster needed for RAC in a Solaris Environment?
Does anyone know, when OCFS will be available for Solaris?

You don't need Veritas Cluster File System, but until OCFS comes out for Solaris you need to think about backups. If you've not got a backup solution that can integrate with rman for an SBT device then backups become more tricky.
If you use ASM then you can take a backup to a "cluster filesystem" (although ASM is raw partitions think of it as a cluster filesystem), that both nodes can see. BUT you then need to get these to tape somehow, unless you've got NetBackup et al. that support RMAN and can backup direct to tape you're more stuck.
Too many people don't think about this. You could create an NFS mount and backup to this from the nodes.

Real Application Cluster on Sun Solaris 8 and Sun Cluster 3

Hello,
we want to install Oracle 9i Enterprise Edition in combination with Oracle Real Application Cluster-Option on 2 Nodes. Every node (12-CPU-SMP-Machine) should running Sun Solaris 8 and Sun Cluster 3 Service.
Does this configuration work with ORAC? I found nowhere informations about. Is there anything I have to pay special attention for during installation?
Thank you for helping and best regards from Berlin/Germany
Michael Wuttke

Forms and report services work fine on solaris 8.
My problem is on the client side.
I have to use solaris 8 with netscape like forms clients
and I wasn't able to make it work with java plugins.
Any solution?
Mauro

IDS 5.0 and Sun-Cluster

I want to make IDS 5.0 highly available using SunCluster. I have few questions about it.
1. Can I install IDS on local disk and make it highly availbale. Sun-Cluster doc says it should be on shared disk.
2. I have an already installed IDS I want to make it highly available using Sun-Cluster, what steps I should follow to achive this.

I suppose that the answer is that it is not a simple task and it depends on the kind of cluster you want to deploy.
I suggest that you carefully read the documentation of Sun Cluster and specifically the Directory Server specific parts.
The way to do it is different with Sun Cluster 2 and Sun Cluster 3.0....
Or you can request help from Sun Professional Services...
Regards,
Ludovic.

Solaris MC and Sun Cluster?

Could someone tell me the relationship between Solaris MC and Sun Cluster?
Thanks!
Bill

Solaris MC is a distributed operating system for
multi-computers (i.e., clusters of nodes) that provides a
single-system image: a cluster appears to the user and applications
as a single computer running the SolarisTM operating system. Solaris
MC is built as a set of extensions to the base Solaris UNIX � system
and provides the same ABI/API as Solaris, running unmodified
applications. The components of Solaris MC are implemented in C++
through a CORBA-compliant object-oriented system with all new services
defined by the IDL definition language. Objects communicate through a
runtime system that borrows from Solaris doors and Spring
subcontracts. Solaris MC is designed for high availability: if a node
fails, the remaining nodes remain operational. Solaris MC has a
distributed caching file system with UNIX consistency semantics,
based on the Spring virtual memory and file system architecture.
Process operations are extended across the cluster, including remote
process execution and a global /proc file system. The external network
is transparently accessible from any node in the cluster.

Sun Cluster & 6130/6140 thru switch with cross-connections not supported?

Hi:
I noticed that the 6140 does not support cross-connecting the 2 controllers to 2 switches for higher availability when using Sun Cluster:
http://docs.sun.com/source/819-7497-10/chapter3.html
Does anyone know why this restriction is there?
Thanks!

Since there was no real answer to the question in this forum, I cross posted this issue to the cluster forum.
See http://forum.java.sun.com/thread.jspa?threadID=5261282&tstart=0 for the full thread.
Basically, the restriction against cross-connections is no longer valid and the documentation should be updated to remove the note.
This is all a good thing, because I had my 6140's wired into my sun cluster environment via the 'cross-connections' method diagramed in figure 3-4. :-)

Failover Zones / Containers with Sun Cluster Geographic Edition and AVS

Hi everyone,
Is the following solution supported/certified by Oracle/Sun? I did find some docs saying it is but cannot find concrete technical information yet...
* Two sites with a 2-node cluster in each site
* 2x Failover containers/zones that are part of the two protection groups (1x group for SAP, other group for 3rd party application)
* Sun Cluster 3.2 and Geographic Edition 3.2 with Availability Suite for SYNC/ASYNC replication over TCP/IP between the two sites
The Zones and their application need to be able to failover between the two sites.
Thanks!
Wim Olivier

Fritz,
Obviously, my colleagues and I, in the Geo Cluster group build and test Geo clusters all the time :-)
We have certainly built and tested Oracle (non-RAC) configurations on AVS. One issue you do have, unfortunately, is that of zones plus AVS (see my Blueprint for more details http://wikis.sun.com/display/BluePrints/Using+Solaris+Cluster+and+Sun+Cluster+Geographic+Edition). Consequently, you can't built the configuration you described. The alternative is to sacrifice zones for now and wait for the fixes to RG affinities (no idea on the schedule for this feature) or find another way to do this - probably hand crafted.
If you follow the OHAC pages (http://www.opensolaris.org/os/community/ha-clusters/) and look at the endorsed projects you'll see that there is a Script Based Plug-in on the way (for OHACGE) that I'm writing. So, if you are interested in playing with OHACGE source or the SCXGE binaries, you might see that appear at some point. Of course, these aren't supported solutions though.
Regards,
Tim
---

Sun cluster and WS7

Where I can find documentation about ws7 and sun cluster?
Anyone has done this with version 7.?

sun cluster shipped with sun java enterprise system 5 does support web server 7. you can get JES5 bits from
http://www.sun.com/software/swportfolio/get.jsp?71220fd6-8b25-4297-8bb7-2bb4c2605e61=1
and more information about JES from
http://www.sun.com/software/javaenterprisesystem/index.jsp

Sun Cluster.. Why?

What are the advantages of installing RAC 10.2.0.3 on a Sun Cluster.? Are there any benefits?

Oracle 10g onward, there is no such burning requirement for Sun Cluster (or any third party cluster) as far as you are using all Oracle technologies for your Oracle RAC database. You should Oracle RAC with ASM for shared storage and that would not require any third party cluster. Bear inmind that
You may need to install Sun Cluster in the following scenarios:
1) If there is applicaiton running with in the cluster along with Oracle RAC database that you want to configure for HA and Sun Cluster provide the cluster resourced (easy to use) to manage and monitor the application. THIS can be achieved with Oracle Clusterware but you will have to write your own cluster resource for that.
2) If you want to install cluster file system such as QFS then you will need to install the Sun Cluster. If this cluster is only running the Oracle RAC database then you can rely on Oracle technologies such as ASM, raw devices without installing Sun Cluster.
3) Any certification conflicts.
Any correction is welcome..
-Harish Kumar Kalra

Multiple Oracle databases in Sun cluster 3.2 (without RAC setup)

There are 2 Sun SPARC (Sun Fire T2000) servers with Solaris 10 (05/09) OS and Sun Cluster 3.2 software. We need two different Oracle databases & instances (Oracle 10g R2 without RAC) for an application. The first database is for Production and need to be configured in the first node & Shared storage disk and need high availability. This database should run from the second node if the first node fails. The second database is for Quality/Test and it is prefered to be running in the Second node for better load distribution. This DB doesnt require any failover.
The shared storage is Sun SE 3510 FC and multiple LUNs can be created for different databases..
Is it possible to configure two different resource groups (one for Quality and other for Production) and make the first node primary for Production RG, and the second node primary for Quality RG and thus distributing the load on 2 servers ? If possible, what special configuration required in Sun OS and Cluster side ?
Appreciate if you can give some configuration procedures/documents for this multi-master cluster setup..

You can configure two resource groups, such as:
# clrg create -n node-a,node-b prod-rg
# clrg create -n node-b qa-rgand you configure the required resources (disk groups / file systems, logical host, oracle listener, oracle server) as described within
[http://docs.sun.com/app/docs/doc/819-2980?l=en&a=expand|http://docs.sun.com/app/docs/doc/819-2980?l=en&a=expand]
Note that this is not really called a "multi-master" configuration - that has a specific meaning for a resource group (see [http://docs.sun.com/app/docs/doc/820-4682/babefcja?l=en&a=view|http://docs.sun.com/app/docs/doc/820-4682/babefcja?l=en&a=view] ) for details.
With Solaris Cluster all nodes part of a cluster are considered active and can host resource groups. You can have any number of resource groups running, where a subset runs one one node, another subset on other nodes. The nodelist property of the resource group defines where it can run, the first node in the list is the preferred primary.
You can even define resource group dependencies or affinities between the resource groups. Like you could define a negative affinity between qa-rg and prod-rg, such as if prod-rg needs to failover to node-b (since e.g. node-a died), it would offline qa-rg. Details for that kind of possibilities are described at [http://docs.sun.com/app/docs/doc/820-4682/ch14_resources_admin-35?l=en&a=view|http://docs.sun.com/app/docs/doc/820-4682/ch14_resources_admin-35?l=en&a=view].
Regards
Thorsten

Sun cluster failed when switching, mount /global/ I/O error .

Hi all,
I am having a problem during switching two Sun Cluster nodes.
Environment:
Two nodes with Solaris 8 (Generic_117350-27), 2 Sun D2 arrays & Vxvm 3.2 and Sun Cluster 3.0.
Porblem description:
scswitch failed , then scshutdown and boot up the both nodes. One node failed because of vxvm boot failure.
The other node is booting up normally but cannot mount /global directories. Manually mount is working fine.
# mount /global/stripe01
mount: I/O error
mount: cannot mount /dev/vx/dsk/globdg/stripe-vol01
# vxdg import globdg
# vxvol -g globdg startall
# mount /dev/vx/dsk/globdg/mirror-vol03 /mnt
# echo $?
0
port:root:/global/.devices/node@1/dev/vx/dsk 169# mount /global/stripe01
mount: I/O error
mount: cannot mount /dev/vx/dsk/globdg/stripe-vol01
Need help urgently
Jeff

I would check your patch levels. I seem to remember there was a linker patch that cause an issue with mounting /global/.devices/node@X
Tim
---

SUn CLuster 3.2 install - scinstall on x86 32 bit Solaris 5.10

Ok - I have 2 machines with 32-bit x86 SOlaris 5.10 - I installed the Sun Cluster 3.2 software but everytime I try scinstall it says rebooting other node and the other node never brings sun cluster up -
Questions:
1. The private interface does not come up on reboot - should it - and should I have an entry for cluster-priv1 in /etc/hosts
2. I know I am supposed to do a scvx -xv - I then try to enable the cluster service, but everything says disabled
I have tried this 5 times no luck - I have lots of cluster experience - and can get Oracle CRS working fine
Any thoughts

Yeahh, guys!!!
I was trying to establish a two-node cluster using VirtualBox + Solaris x86 + Sun Cluster 3.2. The node where I was running scinstall to configure my cluster environment was rebooting the other node in the end of the configuration process but it was hanging in the "Rebooting node01..." message just because it was not able to establish the cluster.
After see your comments, I changed Solaris x86 to Solaris Express Community Edition and Sun Cluster to Cluster Express and now everything is working fine!
Thanks!
Jansen Sena <[email protected]>

Sun cluster, 6140 and 'cross-connections'

This was brought up in the storage forum by somebody else, but the responses never answered the original question:
In the 6140 setup document located at http://docs.sun.com/source/819-7497-11/chapter3.html#50589714_93886 it shows two different ways to cable a host to a 6140 via a SAN switch. figure 3-3 and 3-4
It states that the setup in 3-4 is not supported in a sun cluster environment.
The problem is that given the acive/passive nature of the 6140, the setup shown in 3-4 is the obvious one to use since it prevents one from having to force all of the luns over in the event of an hba port or switch failure.
To make life more interesting, the 6140 setup doc does not make not of what version of sun cluster it is not supported on. Or what the bugid is, or any informationto know if the restriction is still valid.
So, does this restriction still exist? If so, for what version of sun cluster? What version of solaris?

Thanks for the clarification.
As an aside, it would be nice if in the future, the documentation could contain a bit more information than just a simple note saying 'this is not supported'. A reference to a bugid or info doc would go a long way in helping folks determine if the restriction is still valid.
--john

Oracle recommend configuring a Sun StorageTek 6140 with 15 hard disks

Hi,
How would Oracle recommend configuring a Sun StorageTek 6140 with 15 hard disks for optimal use with multiple Solaris servers running Oracle instances connected via a SAN? Should the storage array be configured as a single RAID 5 device and then LUNs created for the different servers? Or should each Oracle instance have its own dedicated hard disks?
Also might we see better performance if we used Solid State Devices for the ZFS Intent Log (ZIL) and/or L2ARC on ZFS, instead of using the UFS file system straight to the SAN?
Regards
NM

I don't think Oracle would recommended any particular way. I would consider the following, but your mileage may vary, so testing is very important.
Raid 1 (10) for logs
RAID5 for Tables
But if you have many servers access the same LUN's/VDISKS then contention may be a problem. Maybe consider consolidating all you servers to a single Oracle Server.( For protection a second server with Oracle Solaris Cluster)
ASM is the best method for managing disks and data placement. But if like to see files/directories then go for UFS with direct I/O enabled. ZFS is fantastic if you want to use Snapshot/clones/Compression, etc.. but I think UFS is faster.
As for cache, Oracle 11gR2 has support for storing objects specifically in FLASH, so look to use the Sun F20 card with 96Gb of flash.
HTH
Andy

What is the best Practice to improve MDIS performance in setting up file aggregation and chunk size

Hello Experts,
in our project we have planned to do some parameter change to improve the MDIS performance and want to know the best practice in setting up file aggregation and chunk size when we importing large numbers of small files(one file contains one record and each file size would be 2 to 3KB) through automatic import process,
below is the current setting in production:-
Chunk Size=2000
No. Of Chunks Processed In Parallel=40
file aggregation-5
Records Per Minute processed-37
and we made the below setting in Development system:-
Chunk Size=70000
No. Of Chunks Processed In Parallel=40
file aggregation-25
Records Per Minute processed-111
after making the above changes import process improved but we want to get expert opinion making these changes in production because there is huge number different between what is there in prod and what change we made in Dev.
thanks in advance,
Regards
Ajay

Hi Ajay,
The SAP default values are as below
Chunk Size=50000
No of Chunks processed in parallel = 5
File aggregation: Depends largely on the data , if you have one or 2 records being sent at a time then it is better to cluster them together and send it at one shot , instead of sending the one record at a time.
Records per minute Processed - Same as above
Regards,
Vag Vignesh Shenoy

Storagetek 6140 - chunk size? - veritas and sun cluster tuning?

Similar Messages

Maybe you are looking for